system

Abstract

We provide the system. [Solution] A means for acquiring a dataset containing information specific to a particular region through data collection methods, The analysis means includes means for identifying biases included in the dataset, A feedback mechanism is designed, and means are used to perform evaluations that take into account the diversity of specific regions in order to mitigate the bias. The dissemination of materials provides a means to provide standards for reducing bias, A means by which a review body evaluates appropriate use based on bias reduction criteria and issues a certification mark, A system that includes this.

Description

Technical Field

[0001] The technology of the present disclosure relates to a system.

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, and includes steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a character of the chatbot, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance that responds to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In conventional artificial intelligence technologies, there is a problem that the use of learning data without considering the cultural and social diversity of a specific region causes bias. This bias may lead to inaccurate or inappropriate results in the judgment results and products of artificial intelligence, and may damage social trust. The purpose of the present invention is to reduce these biases and promote the fair and ethical use of artificial intelligence in a specific region.

Means for Solving the Problems

[0005] This invention is characterized by acquiring a dataset containing region-specific information and identifying bias using analytical means. Furthermore, it designs a feedback mechanism to perform evaluation and authentication to mitigate the aforementioned bias. This supports the fair and ethical dissemination of artificial intelligence by assigning authentication marks to appropriate users and providing information that improves reliability.

[0006] "Data collection means" refers to technology or equipment for acquiring datasets that contain information specific to a particular region.

[0007] "Analysis means" refers to an algorithm or device for identifying biases contained in the acquired dataset.

[0008] A "feedback mechanism" is a structure or process for conducting evaluations that take into account the diversity of a specific region in order to reduce bias.

[0009] "Dissemination materials" are documents or tools that provide companies and organizations with standards for reducing bias.

[0010] A "certification body" is a system or process that is responsible for evaluating appropriate use based on bias reduction criteria and issuing certification marks.

[0011] A "certification mark" is a mark or symbol that serves as proof of trust, given to users who meet bias reduction criteria.

[0012] A "user interface" is a mechanism provided through screens and input devices for collecting feedback.

[0013] "Information that promotes trustworthiness" refers to data and messages provided to authenticated and legitimate users to enhance their trustworthiness. [Brief explanation of the drawing]

[0014] [Figure 1]It is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] It is a conceptual diagram showing an example of the main functions of a data processing device and a smart device according to the first embodiment. [Figure 3] It is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] It is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] It is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] It is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7] It is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] It is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] It shows an emotion map to which a plurality of emotions are mapped. [Figure 10] It shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

MODE FOR CARRYING OUT THE INVENTION

[0015] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0016] First, the terms used in the following description will be explained.

[0017] In the following embodiments, the numbered processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Also, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), an APU (Accelerated Processing Unit), and the like.

[0018] In the following embodiments, the numbered RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a work memory by the processor.

[0019] In the following embodiments, the numbered storage is one or more non-volatile storage devices that store various programs and various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes, etc.

[0020] In the following embodiments, the numbered communication I / F (Interface) is an interface including a communication processor and an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applicable to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark), etc.

[0021] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0022] [First Embodiment]

[0023] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0024] As shown in Figure 1, the data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0025] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0026] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0027] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0028] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0029] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0030] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0031] As shown in Figure 2, in the data processing device 12, a specific processing is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" related to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 according to the specific processing program 56 executed on the RAM 30.

[0032] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0033] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0034] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0035] This invention relates to a method and apparatus for constructing and operating an artificial intelligence system that takes into account the specific information of a particular region. This system consists of a server, a terminal, and a user, and each component works in cooperation with the others.

[0036] First, the server collects data from various companies and organizations. This data includes social and cultural information specific to a particular region and is collected using data collection methods. Next, the server uses analytical tools to identify biases within the data. During the analysis process, statistical and machine learning algorithms are used to find specific patterns and correlations, with a particular focus on bias analysis.

[0037] Based on the analysis results, a feedback mechanism is designed. This mechanism collects feedback from users via a terminal and works in conjunction with the analysis means to evaluate bias reduction. The terminal provides an interface that allows users to easily provide feedback. The collected feedback is then used for further analysis on the server.

[0038] Furthermore, the server uses readily available resources to provide companies with best practices and guidelines to mitigate bias. Based on this information, terminals can improve their AI models and take measures to enhance fairness.

[0039] Furthermore, companies that have successfully implemented appropriate bias reduction measures will be granted a certification mark by the certification body. This occurs after users evaluate their own AI systems, and the server assesses whether the system meets the certification standards based on that evaluation. This certification mark serves as a symbol for users to demonstrate their trustworthiness to consumers.

[0040] To give a concrete example, suppose a user provides past misrecognition data regarding specific attributes to mitigate bias in image recognition AI within a particular cultural sphere. Through this feedback, the device updates the model, the server verifies that the improvements are appropriate, and then issues a certification mark. This process is expected to improve the fairness and reliability of the AI.

[0041] The following describes the processing flow.

[0042] Step 1:

[0043] The server collects region-specific information from various companies and organizations. The server uses data collection methods to obtain cultural and social datasets for the specific region.

[0044] Step 2:

[0045] The server processes the collected dataset using analytical tools to identify biases and patterns. The server applies machine learning algorithms for bias analysis to recognize specific biases and imbalances.

[0046] Step 3:

[0047] The server designs a feedback mechanism based on the analysis results. The terminal builds an interface that allows users to easily input feedback.

[0048] Step 4:

[0049] Users provide feedback on the AI ​​system's output via their devices. Users comment on and evaluate specific patterns or questionable results.

[0050] Step 5:

[0051] The server re-evaluates the collected feedback using analytical tools. The server recalculates the degree of bias using the feedback data and identifies areas for improvement in the AI ​​model.

[0052] Step 6:

[0053] The device provides best practices and guidelines for reducing bias using dissemination materials. The device distributes instructional information to users for improvement.

[0054] Step 7:

[0055] Users update or improve their AI models according to the provided guidelines. Users implement specific changes to reduce bias.

[0056] Step 8:

[0057] The server evaluates the improved model with a review body and determines whether it is suitable for certification. The server performs the evaluation based on the certification standards and records the results.

[0058] Step 9:

[0059] The server assigns a certification mark to companies that it deems to have adequately reduced bias. The server then notifies the user of the result and provides it as reliability information.

[0060] (Example 1)

[0061] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0062] Artificial intelligence systems that take into account the cultural diversity of specific domains are highly susceptible to bias. Therefore, there is a need to build fair and reliable systems for specific regions and situations, but current technology makes this difficult to achieve. Furthermore, there is a lack of evaluation criteria to mitigate bias and methods to promote reliability, necessitating a comprehensive solution.

[0063] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0064] In this invention, the server includes means for acquiring an information set containing information specific to a particular domain using an information gathering device, means for identifying biases included in the information set using an analysis device, and means for collecting opinions from users via voice and text input through a dialogue device and performing evaluations for bias reduction. This makes it possible to efficiently collect information specific to a particular domain, identify biases, and appropriately evaluate and reduce them.

[0065] An "information gathering device" is a mechanism for acquiring data, including cultural and social information, related to a specific domain.

[0066] An "information set" is a collection of cultural and social data related to a specific domain, acquired by an information gathering device.

[0067] An "analytical device" is a technical device used to analyze data contained in an information set and identify bias.

[0068] "Bias" refers to an imbalance or bias present in data, indicating a lack of fairness regarding specific attributes or patterns.

[0069] A "dialogue device" is a mechanism that collects opinions and feedback from users through voice or text input and provides data for bias evaluation.

[0070] "Users" refer to those who provide opinions and feedback through dialogue devices and are actively involved in the biased evaluation process.

[0071] "Dissemination materials" are informational materials that present best practices and guidelines for reducing bias and are intended to be disseminated to users.

[0072] "Guidelines" refer to standards and procedures designed to mitigate bias and maintain and improve the fairness of the system.

[0073] A "review body" refers to an organization or process that evaluates the appropriate use of a service in relation to the goal of reducing bias and issues approval marks.

[0074] An "approval mark" indicates that a product has been certified by an accreditation body as being used appropriately, and serves as proof of its reliability and fairness.

[0075] One embodiment of the present invention is to consider cultural and social information related to a specific domain, mitigate bias through an artificial intelligence system, and achieve fair data processing. This system consists of a server, terminals, and users working together.

[0076] The server first uses an information gathering device to obtain specific information about a particular domain from various sources. This device plays a role in continuously collecting data through APIs and external database connections. The server can efficiently acquire data using the Python "requests" library.

[0077] The acquired information is stored as an information set and analyzed via an analysis device on the server. The analysis includes data preprocessing and bias identification, utilizing statistical analysis tools such as "pandas" and "scikit-learn," as well as machine learning libraries. This makes it possible to detect imbalanced patterns within the data and identify bias.

[0078] Users provide opinions based on their everyday experiences and observations using a dialogue device via their terminal. This feedback is submitted via voice or text input and sent to the server through review forms or mobile apps. The terminal then passes the collected opinions to an analysis device to help evaluate and mitigate bias.

[0079] Based on the collected and analyzed data, the server generates guidelines and best practices for bias reduction as dissemination materials. These materials are provided through a portal site accessible to companies and organizations.

[0080] An example of a feedback prompt is, "Please provide your feedback on how we can improve the misrecognition of image recognition AI in a specific social group." This prompt elicits useful opinions based on users' real-world experiences and is used for advanced data analysis.

[0081] Furthermore, if proper use is confirmed, the server will be granted an approval mark through a review mechanism and will transmit information to promote its reliability. This will enable the system to be positioned as fair and reliable for consumers and related organizations.

[0082] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0083] Step 1: Data Collection

[0084] The server utilizes information gathering devices to collect data related to a specific domain. Inputs include information from APIs and external databases, and the server retrieves this data using libraries such as "requests". The output is a dataset encompassing cultural and social information within the specific domain. In this step, a script is executed periodically to retrieve the latest data.

[0085] Step 2: Data Analysis

[0086] The server analyzes the collected dataset using an analytical instrument. The input for this step is the dataset collected in step 1. The server uses the "pandas" and "scikit-learn" libraries to identify patterns and biases in the data. By applying statistical methods, it is possible to verify the integrity of the data and identify biases. The output is an analysis result that identifies biases.

[0087] Step 3: Gathering Feedback

[0088] The terminal uses a dialogue device to collect opinions from the user. The input consists of user opinions in both voice and text format. The terminal collects this feedback through forms and applications and sends it to the server. The output is the collected user feedback data.

[0089] Step 4: Feedback analysis and model update

[0090] The server updates the AI ​​model based on collected feedback to mitigate bias. The input consists of feedback data and previously obtained data analysis results. The generated AI model is retrained using the "TENSORFLOW®" or "PyTorch" framework to update it to a model that takes bias into account. The output is the improved AI model.

[0091] Step 5: Generating and distributing guidelines

[0092] The server generates guidelines for bias reduction based on the analysis results and updated models. The input consists of the analysis results and improvements to the model. The generated guidelines are output as Markdown or PDF files, and this information is distributed to companies and organizations through a dedicated portal.

[0093] Step 6: Evaluation and Authentication

[0094] The server uses an evaluation system to verify the appropriateness of bias mitigation and issues approval marks as needed. Inputs include evaluation criteria, feedback from companies, and analysis reports. After the review process, eligible users can obtain approval marks. Outputs include approval marks and information to promote reliability.

[0095] (Application Example 1)

[0096] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0097] There is a need for methods to mitigate bias in artificial intelligence systems that take into account regionally differing social and cultural backgrounds, thereby improving fairness and reliability. In particular, in technologies closely related to daily life, such as household robots, there is a need to improve the situation as neglecting regional characteristics can lead to model malfunctions and a decline in the quality of the user experience.

[0098] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0099] In this invention, the server includes means for collecting a data set containing attribute information of a specific region using a data acquisition device, means for identifying biases included in the data set using an analysis device, and means for collecting feedback through a user interface and evaluating bias correction that takes into account the diversity of the specific region. This makes it possible to provide an impartial and reliable artificial intelligence system adapted to a specific region.

[0100] A "data acquisition device" is a device for efficiently collecting data sets that include attribute information of a specific region.

[0101] An "analysis device" is a device used to identify biases in a collected data set and to analyze patterns within the data using statistical and machine learning algorithms.

[0102] A "user interface" is an interactive means of collecting feedback from users and evaluating bias corrections that take into account the diversity of a specific region.

[0103] A "learning algorithm" is a set of computational methods and programs used to continuously modify and adjust a machine learning model based on collected feedback data.

[0104] A "distribution device" is a device that provides users and related organizations with guidelines and standard information for proper bias correction.

[0105] An "evaluation device" is a device that evaluates proper operation based on bias correction guidelines and assigns a confidence mark if the standards are met.

[0106] This invention constitutes a region-adaptive artificial intelligence system specifically for household robots. The server efficiently collects region-specific attribute information using a data acquisition device. This data reflects the cultural and social background of a particular region, and bias is identified by an analysis device based on the collected data set. Machine learning libraries such as TensorFlow are used to identify bias, and a model is created to mitigate the bias based on the obtained data.

[0107] The user's device collects feedback through the user interface, and this feedback is sent to the server. The server uses this data to modify the model using a learning algorithm. In particular, corrections are made that take into account individual cultural elements and regional characteristics based on the feedback. Analysis libraries such as scikit-learn are used in this process. The corrected model is then applied by the distribution device to provide highly accurate, region-specific services.

[0108] As a concrete example, a robot designed for households in Tokyo could gather feedback from users about their cleaning methods and supplement its knowledge with specialized information on how to care for tatami mats. This would allow users to receive high-quality, locally tailored services.

[0109] As an example of a specific prompt for the generating AI model, input would be: "Generate cleaning guidelines that a household robot in Tokyo should follow. Pay attention to how to care for tatami mats." This sentence will serve as the basis for providing region-specific instructions to the AI ​​model.

[0110] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0111] Step 1:

[0112] The server collects regional attribute information using data acquisition devices. It receives social and cultural data related to a specific region as input and generates a data set containing regional characteristics as output. Specifically, it aggregates data provided by local businesses and publicly available statistical data.

[0113] Step 2:

[0114] The server identifies bias within a data set collected using an analysis device. It receives the data set as input and performs data analysis using TensorFlow to identify bias. The output generates a report detailing the degree of bias and the types of bias identified. Specific operations include feature extraction and correlation analysis.

[0115] Step 3:

[0116] The user's device collects feedback using a user interface. It acquires feedback data based on user opinions and improvement requests as input. It creates a feedback dataset to be sent to the server as output. Specific operations include data input in the form of questionnaires and opinion collection using speech recognition.

[0117] Step 4:

[0118] The server modifies the machine learning model using a learning algorithm based on feedback data. It receives bias analysis results and user feedback data as input and generates a modified AI model as output. Specifically, it performs data retraining and model parameter adjustment using scikit-learn.

[0119] Step 5:

[0120] The server distributes the corrected model to the terminal via a distribution device and applies it. It receives the corrected AI model as input and reflects the optimized model in the terminal as output. Specifically, it provides operational guidelines based on newly learned processes and updates the model within the terminal.

[0121] Step 6:

[0122] Users will experience region-specific services suggested by robots based on improved AI models and provide feedback on their evaluation. The expected output is improved efficiency of region-specific robot services and increased user satisfaction. Specific examples of operation include the robot suggesting methods for maintaining tatami mats and the user verifying the results after application.

[0123] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0124] This invention relates to a method and apparatus for enhancing the bias reduction process in an artificial intelligence system that takes into account specific regional information, by also analyzing the user's emotions. The system consists of a server, a terminal, a user, and an emotion engine.

[0125] First, the server collects data provided by various companies and organizations. This includes cultural and social information specific to a particular region, and is efficiently acquired through data collection methods. Next, the server uses analysis tools to identify biases within the data and detect specific patterns. In this process, in addition to traditional machine learning algorithms, an emotion engine is used to analyze the sentiment of feedback collected from users.

[0126] The emotion engine operates through a process of processing user feedback and analyzing the emotions it captures. For example, if a user expresses dissatisfaction with an image recognition result, the emotion engine examines the cause of that emotion, and this is taken into consideration in the feedback mechanism. This leads to more precise recognition of biases, and the feedback mechanism is used to implement improvements.

[0127] The terminal provides an interface for users to input feedback. The feedback entered through the user interface is analyzed by an emotion engine, sent to a server, and used for bias assessment and improvement.

[0128] Furthermore, the server provides guidelines and tools to mitigate bias using readily available materials. This information is distributed to companies via terminals, promoting the proper improvement of AI models. In addition, a review body evaluates the improved AI models and issues a certification mark if they meet the standards.

[0129] A concrete example is an AI that generates marketing images for consumers in a specific region. This system evaluates both the user's feedback regarding bias in the image content and the user's emotions simultaneously. The server uses this information to adjust the image generation algorithm, producing more culturally appropriate output. As a result, user satisfaction is expected to improve, and trust will be strengthened.

[0130] The following describes the processing flow.

[0131] Step 1:

[0132] The server collects data provided by companies and organizations through data collection methods. This includes datasets that contain cultural background and social information specific to a particular region.

[0133] Step 2:

[0134] The server processes the collected data using analytical tools to identify any biases or patterns present in the data. The analysis results are recorded in a database on the server.

[0135] Step 3:

[0136] Users input feedback on the AI ​​system's output using the terminal's user interface. This feedback includes specific comments and suggestions for improvement.

[0137] Step 4:

[0138] The device sends user feedback to the emotion engine. The emotion engine analyzes the user's emotions contained in the feedback and determines the type and intensity of those emotions.

[0139] Step 5:

[0140] The server adjusts the feedback mechanism using the analysis results from the emotion engine. In particular, it evaluates areas where strong biases were identified and identifies necessary improvements.

[0141] Step 6:

[0142] The device receives sentiment analysis results from the server and presents the user with improvement measures and suggestions based on the feedback. It also provides guidelines for reducing bias through promotional materials.

[0143] Step 7:

[0144] Users review the suggested improvements via their devices and implement them into their own AI model. They adjust the model parameters as needed.

[0145] Step 8:

[0146] The server activates a review mechanism to evaluate the AI ​​model improved by the user. The evaluation is conducted based on bias mitigation criteria, and if it passes, a certification mark is issued.

[0147] Step 9:

[0148] The device notifies the user that it has been assigned an authentication mark and provides information informing them of the improved reliability of the authenticated AI system.

[0149] (Example 2)

[0150] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0151] There is a need to effectively identify and mitigate biases in datasets within specific regions. However, conventional technologies struggle to adequately consider regional cultural and emotional factors in their bias mitigation. Furthermore, the lack of mechanisms for effectively incorporating user feedback limits model improvement.

[0152] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0153] In this invention, the server includes means for acquiring an information set containing information about local culture and society using a data processing device, means for detecting biases contained in the information set using an analysis device, and means for identifying the emotions of feedback provided by users using an emotion analysis device and reflecting this in the bias detection results. This makes it possible to make appropriate improvements that take into account region-specific biases.

[0154] A "data processing device" is a device for efficiently collecting information about local culture and society and obtaining it as an information set.

[0155] An "analytical device" is a device that detects data contained in an information set and performs analysis to identify biases.

[0156] An "emotion analysis device" is a device that examines the emotions in feedback provided by users and incorporates them into the bias detection process.

[0157] A "user interface" is a system component that provides an interactive means for users to provide feedback.

[0158] An "information distribution device" is a device that generates standards and tools for bias correction and distributes information in order to disseminate them.

[0159] An "evaluation device" is a device that examines whether an improved information processing method conforms to the standards and, if it does, issues a certification mark.

[0160] A "certification mark" is a mark that indicates reliability, given to an information processing method or user that conforms to standards.

[0161] In this invention, a data processing system is constructed through the cooperation of a server, a terminal, and an emotion analysis engine. The main components include a data processing device, an analysis device, an emotion analysis device, a user interface, an information distribution device, and an evaluation device.

[0162] The server uses data processing equipment to collect local cultural and social information and obtain a collection of that information. This allows the server to accumulate diverse information and form a foundation for mitigating bias. Next, the server uses analytical equipment to detect biases in the collected information collection and extract specific patterns and trends. This analysis takes into account the cultural and social context of the specific region to minimize errors.

[0163] The sentiment analyzer is responsible for processing user feedback and analyzing the emotions associated with it. When a user inputs feedback using a terminal, the sentiment analyzer emotionally evaluates the content and incorporates this into bias detection. This analysis is performed to comprehensively understand user satisfaction and dissatisfaction. The user interface provides users with intuitive and easy-to-use means for collecting and inputting feedback.

[0164] Furthermore, the server uses an information distribution device to generate guidelines and tools for reducing bias and distribute them to companies and related organizations. This distribution supports the dissemination of improved information processing methods. The evaluation device checks whether the improved information processing methods conform to predetermined standards. If they conform to the standards, the server assigns a certification mark, indicating improved reliability.

[0165] A concrete example is an AI system that generates marketing images targeted at consumers in a specific region. If a user provides feedback on biases related to an image, the prompt might be something like, "Please enter information to identify and improve biases in marketing images for a specific region." The server can then use this information to adjust its algorithm and obtain a more culturally appropriate generating AI model. As a result, user satisfaction and reliability are expected to improve.

[0166] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0167] Step 1:

[0168] The server uses data processing equipment to collect local cultural and social information. The input is diverse information provided by companies and organizations, and the output is a collection of that information. In this step, the server retrieves data directly from the information sources and stores it in a database, preparing the foundational data necessary for the next analysis.

[0169] Step 2:

[0170] The server analyzes the data set collected using the analysis device. The input is the data set obtained in step 1, and the output is the identification result that detects and identifies bias. Here, the server applies a machine learning algorithm to extract anti-biased patterns in the data. This analysis is performed using statistical methods.

[0171] Step 3:

[0172] The user provides feedback using the device's user interface. The input is the user's feedback, and the output is the feedback data. In this step, the device's interface provides a user-friendly feedback form, allowing users to easily input their opinions and feelings.

[0173] Step 4:

[0174] The emotion analysis device analyzes the feedback received from the user and performs an emotional evaluation. The input is the feedback data obtained in step 3, and the output is the emotional evaluation result. The emotion analysis device uses natural language processing technology to emotionally interpret the feedback and sends the evaluation result as data to the server.

[0175] Step 5:

[0176] The server re-evaluates the bias in the information set based on the data obtained from sentiment analysis. The input is the bias identification result from step 2 and the sentiment evaluation result from step 4, and the output is the re-evaluated bias data. Here, the server combines the sentiment data with the existing bias data and performs bias adjustment. The sentiment element of feedback is incorporated into the re-evaluation process.

[0177] Step 6:

[0178] The server uses an information distribution device to create guidelines for bias reduction and distributes them to companies and related organizations. The input is re-evaluated bias data, and the output is specific improvement guidelines and tools. The server distributes this through digital media, providing the necessary guidance to widely promote the improvement of AI models.

[0179] Step 7:

[0180] The evaluation device checks whether the improved AI model meets the standards. The input is the improved model and its usage results, and the output is the certification result of compliance with the standards. Here, it is confirmed whether the model's accuracy and reliability meet the standards, and if it does, a certification mark is issued. This evaluation ensures that the model can be used with confidence.

[0181] (Application Example 2)

[0182] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0183] There is a need to consider the diverse cultural backgrounds of specific regions and mitigate bias in content such as advertising. However, conventional systems have struggled to adequately consider regional emotional factors, making it difficult to improve user satisfaction. Furthermore, they lacked the ability to dynamically adjust content, making immediate responses difficult.

[0184] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0185] In this invention, the server includes means for acquiring a dataset containing specific regional information using data collection means, means for identifying biases contained in the dataset using analysis means, means for analyzing user sentiment data using sentiment analysis means and using the results to mitigate bias, and means for dynamically changing content based on the analysis results using dynamic adjustment means. This makes it possible to mitigate bias in a way that is appropriate for the local culture and to adjust content in real time according to the user's emotions.

[0186] "Data collection means" refers to devices and methods for efficiently acquiring datasets containing cultural and social information of a specific region.

[0187] "Analysis means" refers to processes or devices for identifying biases within an acquired dataset and detecting those patterns.

[0188] A "feedback mechanism" is a device or method designed to collect user feedback and use that feedback to reduce bias.

[0189] "Dissemination materials" are means of providing and spreading standards and tools to reduce bias.

[0190] A "certification body" refers to a device or method for evaluating and certifying appropriate use based on bias mitigation criteria.

[0191] "Sentiment analysis methods" are techniques for analyzing user feedback and emotional data and using the results to reduce bias.

[0192] A "dynamic adjustment method" is a means of dynamically adjusting specific content to match user emotions and local culture based on analysis results.

[0193] This invention is a system that takes into account the cultural background of a specific region and the emotions of users to reduce bias in advertising content and improve the user experience. The system mainly consists of data collection means, analysis means, emotion analysis means, dynamic adjustment means, feedback mechanism, and review mechanism.

[0194] The server first uses data collection tools to acquire a dataset containing cultural and social information about a specific region. This dataset is collected from open databases on the internet and information provided by companies. The acquired dataset is then analyzed using analytical tools to identify patterns of bias. This process utilizes data analysis platforms and machine learning algorithms.

[0195] Users view advertising content via their devices, and their emotional data is analyzed using sentiment analysis tools. Sentiment analysis APIs such as IBM Watson® and Google® Cloud Natural Language API are used, and the results are reflected in the analysis report. Based on these results, the server uses dynamic adjustment tools to adjust the advertising content in real time. This adjustment is based on information stored in the Firebase database.

[0196] Furthermore, the server provides standards for reducing bias through dissemination materials and sends information to companies that have been given certification marks to promote trustworthiness. The feedback mechanism collects user feedback and uses it in conjunction with sentiment analysis to achieve less biased output throughout the system.

[0197] A concrete example is local event advertisements. If a user sees an ad and shows no interest, the system can identify the reason through sentiment analysis and adjust the ad to, for example, family-friendly events or content that might interest them.

[0198] Example of a prompt:

[0199] "How can we emotionally analyze user feedback on ads they've watched and tailor ad content to fit specific regional cultures?"

[0200] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0201] Step 1:

[0202] The server uses data collection methods to acquire datasets containing information specific to a particular region. It investigates information sources such as open databases on the internet and those provided by companies, and aggregates this specific information. The input is regional cultural and social data, and the output is the storage of this data in a database.

[0203] Step 2:

[0204] The server analyzes the acquired dataset using analytical tools and identifies bias patterns. A machine learning algorithm is applied to each dataset to evaluate the presence of bias. The input is the dataset collected in step 1, and the output is data showing bias patterns. This result is used for adjustment in subsequent processing.

[0205] Step 3:

[0206] Users view advertising content via their devices, and the sentiment analysis system collects and analyzes the user's emotional data in response. Specifically, inputs such as the user's reactions and tone of voice during ad display are passed to the sentiment API and output as an emotion score. This analysis result is data that indicates the user's emotional state.

[0207] Step 4:

[0208] The server uses sentiment analysis results to dynamically adjust advertising content in real time. The input is the sentiment score from step 3, and the output is the generation of appropriate content using a generative AI model, which is then sent to the device. For example, it automatically changes the content to be more engaging.

[0209] Step 5:

[0210] Users provide feedback using their devices, which a feedback mechanism collects and analyzes. The output is the analysis of the feedback input, with suggestions for improvement to reduce bias sent to the server. This output is then used to further refine the system's recognition algorithm.

[0211] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0212] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0213] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0214] [Second Embodiment]

[0215] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0216] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0217] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0218] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0219] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0220] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0221] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0222] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0223] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0224] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0225] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0226] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0227] This invention relates to a method and apparatus for constructing and operating an artificial intelligence system that takes into account the specific information of a particular region. This system consists of a server, a terminal, and a user, and each component works in cooperation with the others.

[0228] First, the server collects data from various companies and organizations. This data includes social and cultural information specific to a particular region and is collected using data collection methods. Next, the server uses analytical tools to identify biases within the data. During the analysis process, statistical and machine learning algorithms are used to find specific patterns and correlations, with a particular focus on bias analysis.

[0229] Based on the analysis results, a feedback mechanism is designed. This mechanism collects feedback from users via a terminal and works in conjunction with the analysis means to evaluate bias reduction. The terminal provides an interface that allows users to easily provide feedback. The collected feedback is then used for further analysis on the server.

[0230] Furthermore, the server uses readily available resources to provide companies with best practices and guidelines to mitigate bias. Based on this information, terminals can improve their AI models and take measures to enhance fairness.

[0231] Furthermore, companies that have successfully implemented appropriate bias reduction measures will be granted a certification mark by the certification body. This occurs after users evaluate their own AI systems, and the server assesses whether the system meets the certification standards based on that evaluation. This certification mark serves as a symbol for users to demonstrate their trustworthiness to consumers.

[0232] To give a concrete example, suppose a user provides past misrecognition data regarding specific attributes to mitigate bias in image recognition AI within a particular cultural sphere. Through this feedback, the device updates the model, the server verifies that the improvements are appropriate, and then issues a certification mark. This process is expected to improve the fairness and reliability of the AI.

[0233] The following describes the processing flow.

[0234] Step 1:

[0235] The server collects region-specific information from various companies and organizations. The server uses data collection methods to obtain cultural and social datasets for the specific region.

[0236] Step 2:

[0237] The server processes the collected dataset using analytical tools to identify biases and patterns. The server applies machine learning algorithms for bias analysis to recognize specific biases and imbalances.

[0238] Step 3:

[0239] The server designs a feedback mechanism based on the analysis results. The terminal builds an interface that allows users to easily input feedback.

[0240] Step 4:

[0241] Users provide feedback on the AI ​​system's output via their devices. Users comment on and evaluate specific patterns or questionable results.

[0242] Step 5:

[0243] The server re-evaluates the collected feedback using analytical tools. The server recalculates the degree of bias using the feedback data and identifies areas for improvement in the AI ​​model.

[0244] Step 6:

[0245] The device provides best practices and guidelines for reducing bias using dissemination materials. The device distributes instructional information to users for improvement.

[0246] Step 7:

[0247] Users update or improve their AI models according to the provided guidelines. Users implement specific changes to reduce bias.

[0248] Step 8:

[0249] The server evaluates the improved model with a review body and determines whether it is suitable for certification. The server performs the evaluation based on the certification standards and records the results.

[0250] Step 9:

[0251] The server assigns a certification mark to companies that it deems to have adequately reduced bias. The server then notifies the user of the result and provides it as reliability information.

[0252] (Example 1)

[0253] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0254] Artificial intelligence systems that take into account the cultural diversity of specific domains are highly susceptible to bias. Therefore, there is a need to build fair and reliable systems for specific regions and situations, but current technology makes this difficult to achieve. Furthermore, there is a lack of evaluation criteria to mitigate bias and methods to promote reliability, necessitating a comprehensive solution.

[0255] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0256] In this invention, the server includes means for acquiring an information set containing information specific to a particular domain using an information gathering device, means for identifying biases included in the information set using an analysis device, and means for collecting opinions from users via voice and text input through a dialogue device and performing evaluations for bias reduction. This makes it possible to efficiently collect information specific to a particular domain, identify biases, and appropriately evaluate and reduce them.

[0257] An "information gathering device" is a mechanism for acquiring data, including cultural and social information, related to a specific domain.

[0258] An "information set" is a collection of cultural and social data related to a specific domain, acquired by an information gathering device.

[0259] An "analytical device" is a technical device used to analyze data contained in an information set and identify bias.

[0260] "Bias" refers to an imbalance or bias present in data, indicating a lack of fairness regarding specific attributes or patterns.

[0261] A "dialogue device" is a mechanism that collects opinions and feedback from users through voice or text input and provides data for bias evaluation.

[0262] "Users" refer to those who provide opinions and feedback through dialogue devices and are actively involved in the biased evaluation process.

[0263] "Dissemination materials" are informational materials that present best practices and guidelines for reducing bias and are intended to be disseminated to users.

[0264] "Guidelines" refer to standards and procedures designed to mitigate bias and maintain and improve the fairness of the system.

[0265] A "review body" refers to an organization or process that evaluates the appropriate use of a service in relation to the goal of reducing bias and issues approval marks.

[0266] An "approval mark" indicates that a product has been certified by an accreditation body as being used appropriately, and serves as proof of its reliability and fairness.

[0267] One embodiment of the present invention is to consider cultural and social information related to a specific domain, mitigate bias through an artificial intelligence system, and achieve fair data processing. This system consists of a server, terminals, and users working together.

[0268] The server first uses an information gathering device to obtain specific information about a particular domain from various sources. This device plays a role in continuously collecting data through APIs and external database connections. The server can efficiently acquire data using the Python "requests" library.

[0269] The acquired information is stored as an information set and analyzed via an analysis device on the server. The analysis includes data preprocessing and bias identification, utilizing statistical analysis tools such as "pandas" and "scikit-learn," as well as machine learning libraries. This makes it possible to detect imbalanced patterns within the data and identify bias.

[0270] Users provide opinions based on their everyday experiences and observations using a dialogue device via their terminal. This feedback is submitted via voice or text input and sent to the server through review forms or mobile apps. The terminal then passes the collected opinions to an analysis device to help evaluate and mitigate bias.

[0271] Based on the collected and analyzed data, the server generates guidelines and best practices for bias reduction as dissemination materials. These materials are provided through a portal site accessible to companies and organizations.

[0272] An example of a feedback prompt is, "Please provide your feedback on how we can improve the misrecognition of image recognition AI in a specific social group." This prompt elicits useful opinions based on users' real-world experiences and is used for advanced data analysis.

[0273] Furthermore, if proper use is confirmed, the server will be granted an approval mark through a review mechanism and will transmit information to promote its reliability. This will enable the system to be positioned as fair and reliable for consumers and related organizations.

[0274] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0275] Step 1: Data Collection

[0276] The server utilizes information gathering devices to collect data related to a specific domain. Inputs include information from APIs and external databases, and the server retrieves this data using libraries such as "requests". The output is a dataset encompassing cultural and social information within the specific domain. In this step, a script is executed periodically to retrieve the latest data.

[0277] Step 2: Data Analysis

[0278] The server analyzes the collected dataset using an analytical instrument. The input for this step is the dataset collected in step 1. The server uses the "pandas" and "scikit-learn" libraries to identify patterns and biases in the data. By applying statistical methods, it is possible to verify the integrity of the data and identify biases. The output is an analysis result that identifies biases.

[0279] Step 3: Gathering Feedback

[0280] The terminal collects opinions from users using an interactive device. The inputs are the opinions in the form of voice and text by the users. The terminal plays the role of collecting these feedbacks via forms or applications and sending them to the server. As output, the collected user feedback data is generated.

[0281] Step 4: Feedback Analysis and Model Update

[0282] The server updates the AI model based on the collected feedback for bias reduction. The inputs are the feedback data and the previously obtained data analysis results. By retraining the generated AI model using frameworks such as "TensorFlow" or "PyTorch", the model is updated to a bias - considered model. The output is an improved AI model.

[0283] Step 5: Generation and Distribution of Guidelines

[0284] The server generates guidelines for bias reduction based on the analysis results and the updated model. The inputs are the analysis results and the improvement points of the model. The generated guidelines are output as Markdown or PDF, and this information is distributed to companies and organizations through a dedicated portal.

[0285] Step 6: Evaluation and Certification

[0286] The server uses an evaluation system to confirm the appropriateness of bias reduction and grants an approval label if necessary. The inputs are the evaluation criteria, the feedback from the company, and the analysis report. Through the review process, the corresponding users can obtain an approval label. The output is an approval label and information for promoting reliability.

[0287] (Application Example 1)

[0288] Next, Application Example 1 will be described. In the following description, the data processing device 12 is referred to as the "server", and the smart glasses 214 are referred to as the "terminal".

[0289] There is a need for methods to mitigate bias in artificial intelligence systems that take into account regionally differing social and cultural backgrounds, thereby improving fairness and reliability. In particular, in technologies closely related to daily life, such as household robots, there is a need to improve the situation as neglecting regional characteristics can lead to model malfunctions and a decline in the quality of the user experience.

[0290] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0291] In this invention, the server includes means for collecting a data set containing attribute information of a specific region using a data acquisition device, means for identifying biases included in the data set using an analysis device, and means for collecting feedback through a user interface and evaluating bias correction that takes into account the diversity of the specific region. This makes it possible to provide an impartial and reliable artificial intelligence system adapted to a specific region.

[0292] A "data acquisition device" is a device for efficiently collecting data sets that include attribute information of a specific region.

[0293] An "analysis device" is a device used to identify biases in a collected data set and to analyze patterns within the data using statistical and machine learning algorithms.

[0294] A "user interface" is an interactive means of collecting feedback from users and evaluating bias corrections that take into account the diversity of a specific region.

[0295] A "learning algorithm" is a set of computational methods and programs used to continuously modify and adjust a machine learning model based on collected feedback data.

[0296] A "distribution device" is a device that provides users and related organizations with guidelines and standard information for proper bias correction.

[0297] An "evaluation device" is a device that evaluates proper operation based on bias correction guidelines and assigns a confidence mark if the standards are met.

[0298] This invention constitutes a region-adaptive artificial intelligence system specifically for household robots. The server efficiently collects region-specific attribute information using a data acquisition device. This data reflects the cultural and social background of a particular region, and bias is identified by an analysis device based on the collected data set. Machine learning libraries such as TensorFlow are used to identify bias, and a model is created to mitigate the bias based on the obtained data.

[0299] The user's device collects feedback through the user interface, and this feedback is sent to the server. The server uses this data to modify the model using a learning algorithm. In particular, corrections are made that take into account individual cultural elements and regional characteristics based on the feedback. Analysis libraries such as scikit-learn are used in this process. The corrected model is then applied by the distribution device to provide highly accurate, region-specific services.

[0300] As a concrete example, a robot designed for households in Tokyo could gather feedback from users about their cleaning methods and supplement its knowledge with specialized information on how to care for tatami mats. This would allow users to receive high-quality, locally tailored services.

[0301] As an example of a specific prompt for the generating AI model, input would be: "Generate cleaning guidelines that a household robot in Tokyo should follow. Pay attention to how to care for tatami mats." This sentence will serve as the basis for providing region-specific instructions to the AI ​​model.

[0302] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0303] Step 1:

[0304] The server collects the attribute information of the region using a data acquisition device. It receives social and cultural data related to a specific region as input and generates a data set containing regional characteristics as output. Specifically, it aggregates data provided by local companies and published statistical data, etc.

[0305] Step 2:

[0306] The server identifies the biases within the collected data set using an analysis device. It receives the data set as input and performs data analysis to identify biases using TensorFlow. As output, it generates a report on the degree of bias and the types of identified biases. Specific operations include feature extraction and correlation analysis.

[0307] Step 3:

[0308] The user's terminal collects feedback using the user interface. It obtains feedback data based on the user's opinions and improvement requests as input. As output, it creates a feedback data set to be sent to the server. Specific operations include data input in the form of questionnaires and opinion collection through voice recognition.

[0309] Step 4:

[0310] The server modifies the machine learning model using a learning algorithm based on the feedback data. It receives the bias analysis result and user feedback data as input and generates a modified AI model as output. Specifically, data re-learning using scikit-learn and adjustment of model parameters are performed.

[0311] Step 5:

[0312] The server distributes the corrected model to the terminal via a distribution device and applies it. It receives the corrected AI model as input and reflects the optimized model in the terminal as output. Specifically, it provides operational guidelines based on newly learned processes and updates the model within the terminal.

[0313] Step 6:

[0314] Users will experience region-specific services suggested by robots based on improved AI models and provide feedback on their evaluation. The expected output is improved efficiency of region-specific robot services and increased user satisfaction. Specific examples of operation include the robot suggesting methods for maintaining tatami mats and the user verifying the results after application.

[0315] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0316] This invention relates to a method and apparatus for enhancing the bias reduction process in an artificial intelligence system that takes into account specific regional information, by also analyzing the user's emotions. The system consists of a server, a terminal, a user, and an emotion engine.

[0317] First, the server collects data provided by various companies and organizations. This includes cultural and social information specific to a particular region, and is efficiently acquired through data collection methods. Next, the server uses analysis tools to identify biases within the data and detect specific patterns. In this process, in addition to traditional machine learning algorithms, an emotion engine is used to analyze the sentiment of feedback collected from users.

[0318] The emotion engine operates through a process of processing user feedback and analyzing the emotions it captures. For example, if a user expresses dissatisfaction with an image recognition result, the emotion engine examines the cause of that emotion, and this is taken into consideration in the feedback mechanism. This leads to more precise recognition of biases, and the feedback mechanism is used to implement improvements.

[0319] The terminal provides an interface for users to input feedback. The feedback entered through the user interface is analyzed by an emotion engine, sent to a server, and used for bias assessment and improvement.

[0320] Furthermore, the server provides guidelines and tools to mitigate bias using readily available materials. This information is distributed to companies via terminals, promoting the proper improvement of AI models. In addition, a review body evaluates the improved AI models and issues a certification mark if they meet the standards.

[0321] A concrete example is an AI that generates marketing images for consumers in a specific region. This system evaluates both the user's feedback regarding bias in the image content and the user's emotions simultaneously. The server uses this information to adjust the image generation algorithm, producing more culturally appropriate output. As a result, user satisfaction is expected to improve, and trust will be strengthened.

[0322] The following describes the processing flow.

[0323] Step 1:

[0324] The server collects data provided by companies and organizations through data collection methods. This includes datasets that contain cultural background and social information specific to a particular region.

[0325] Step 2:

[0326] The server processes the collected data using analytical tools to identify any biases or patterns present in the data. The analysis results are recorded in a database on the server.

[0327] Step 3:

[0328] Users input feedback on the AI ​​system's output using the terminal's user interface. This feedback includes specific comments and suggestions for improvement.

[0329] Step 4:

[0330] The device sends user feedback to the emotion engine. The emotion engine analyzes the user's emotions contained in the feedback and determines the type and intensity of those emotions.

[0331] Step 5:

[0332] The server adjusts the feedback mechanism using the analysis results from the emotion engine. In particular, it evaluates areas where strong biases were identified and identifies necessary improvements.

[0333] Step 6:

[0334] The device receives sentiment analysis results from the server and presents the user with improvement measures and suggestions based on the feedback. It also provides guidelines for reducing bias through promotional materials.

[0335] Step 7:

[0336] Users review the suggested improvements via their devices and implement them into their own AI model. They adjust the model parameters as needed.

[0337] Step 8:

[0338] The server activates a review mechanism to evaluate the AI ​​model improved by the user. The evaluation is conducted based on bias mitigation criteria, and if it passes, a certification mark is issued.

[0339] Step 9:

[0340] The device notifies the user that it has been assigned an authentication mark and provides information informing them of the improved reliability of the authenticated AI system.

[0341] (Example 2)

[0342] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal".

[0343] There is a need to effectively identify and mitigate biases in datasets within specific regions. However, conventional technologies struggle to adequately consider regional cultural and emotional factors in their bias mitigation. Furthermore, the lack of mechanisms for effectively incorporating user feedback limits model improvement.

[0344] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0345] In this invention, the server includes means for acquiring an information set containing information about local culture and society using a data processing device, means for detecting biases contained in the information set using an analysis device, and means for identifying the emotions of feedback provided by users using an emotion analysis device and reflecting this in the bias detection results. This makes it possible to make appropriate improvements that take into account region-specific biases.

[0346] A "data processing device" is a device for efficiently collecting information about local culture and society and obtaining it as an information set.

[0347] An "analytical device" is a device that detects data contained in an information set and performs analysis to identify biases.

[0348] An "emotion analysis device" is a device that examines the emotions in feedback provided by users and incorporates them into the bias detection process.

[0349] A "user interface" is a system component that provides an interactive means for users to provide feedback.

[0350] An "information distribution device" is a device that generates standards and tools for bias correction and distributes information in order to disseminate them.

[0351] An "evaluation device" is a device that examines whether an improved information processing method conforms to the standards and, if it does, issues a certification mark.

[0352] A "certification mark" is a mark that indicates reliability, given to an information processing method or user that conforms to standards.

[0353] In this invention, a data processing system is constructed through the cooperation of a server, a terminal, and an emotion analysis engine. The main components include a data processing device, an analysis device, an emotion analysis device, a user interface, an information distribution device, and an evaluation device.

[0354] The server uses data processing equipment to collect local cultural and social information and obtain a collection of that information. This allows the server to accumulate diverse information and form a foundation for mitigating bias. Next, the server uses analytical equipment to detect biases in the collected information collection and extract specific patterns and trends. This analysis takes into account the cultural and social context of the specific region to minimize errors.

[0355] The sentiment analyzer is responsible for processing user feedback and analyzing the emotions associated with it. When a user inputs feedback using a terminal, the sentiment analyzer emotionally evaluates the content and incorporates this into bias detection. This analysis is performed to comprehensively understand user satisfaction and dissatisfaction. The user interface provides users with intuitive and easy-to-use means for collecting and inputting feedback.

[0356] Furthermore, the server uses an information distribution device to generate guidelines and tools for reducing bias and distribute them to companies and related organizations. This distribution supports the dissemination of improved information processing methods. The evaluation device checks whether the improved information processing methods conform to predetermined standards. If they conform to the standards, the server assigns a certification mark, indicating improved reliability.

[0357] A concrete example is an AI system that generates marketing images targeted at consumers in a specific region. If a user provides feedback on biases related to an image, the prompt might be something like, "Please enter information to identify and improve biases in marketing images for a specific region." The server can then use this information to adjust its algorithm and obtain a more culturally appropriate generating AI model. As a result, user satisfaction and reliability are expected to improve.

[0358] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0359] Step 1:

[0360] The server uses data processing equipment to collect local cultural and social information. The input is diverse information provided by companies and organizations, and the output is a collection of that information. In this step, the server retrieves data directly from the information sources and stores it in a database, preparing the foundational data necessary for the next analysis.

[0361] Step 2:

[0362] The server analyzes the data set collected using the analysis device. The input is the data set obtained in step 1, and the output is the identification result that detects and identifies bias. Here, the server applies a machine learning algorithm to extract anti-biased patterns in the data. This analysis is performed using statistical methods.

[0363] Step 3:

[0364] The user provides feedback using the device's user interface. The input is the user's feedback, and the output is the feedback data. In this step, the device's interface provides a user-friendly feedback form, allowing users to easily input their opinions and feelings.

[0365] Step 4:

[0366] The emotion analysis device analyzes the feedback received from the user and performs an emotional evaluation. The input is the feedback data obtained in step 3, and the output is the emotional evaluation result. The emotion analysis device uses natural language processing technology to emotionally interpret the feedback and sends the evaluation result as data to the server.

[0367] Step 5:

[0368] The server re-evaluates the bias in the information set based on the data obtained from sentiment analysis. The input is the bias identification result from step 2 and the sentiment evaluation result from step 4, and the output is the re-evaluated bias data. Here, the server combines the sentiment data with the existing bias data and performs bias adjustment. The sentiment element of feedback is incorporated into the re-evaluation process.

[0369] Step 6:

[0370] The server uses an information distribution device to create guidelines for bias reduction and distributes them to companies and related organizations. The input is re-evaluated bias data, and the output is specific improvement guidelines and tools. The server distributes this through digital media, providing the necessary guidance to widely promote the improvement of AI models.

[0371] Step 7:

[0372] The evaluation device checks whether the improved AI model meets the standards. The input is the improved model and its usage results, and the output is the certification result of compliance with the standards. Here, it is confirmed whether the model's accuracy and reliability meet the standards, and if it does, a certification mark is issued. This evaluation ensures that the model can be used with confidence.

[0373] (Application Example 2)

[0374] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0375] There is a need to consider the diverse cultural backgrounds of specific regions and mitigate bias in content such as advertising. However, conventional systems have struggled to adequately consider regional emotional factors, making it difficult to improve user satisfaction. Furthermore, they lacked the ability to dynamically adjust content, making immediate responses difficult.

[0376] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0377] In this invention, the server includes means for acquiring a dataset containing specific regional information using data collection means, means for identifying biases contained in the dataset using analysis means, means for analyzing user sentiment data using sentiment analysis means and using the results to mitigate bias, and means for dynamically changing content based on the analysis results using dynamic adjustment means. This makes it possible to mitigate bias in a way that is appropriate for the local culture and to adjust content in real time according to the user's emotions.

[0378] "Data collection means" refers to devices and methods for efficiently acquiring datasets containing cultural and social information of a specific region.

[0379] "Analysis means" refers to processes or devices for identifying biases within an acquired dataset and detecting those patterns.

[0380] A "feedback mechanism" is a device or method designed to collect user feedback and use that feedback to reduce bias.

[0381] "Dissemination materials" are means of providing and spreading standards and tools to reduce bias.

[0382] A "certification body" refers to a device or method for evaluating and certifying appropriate use based on bias mitigation criteria.

[0383] "Sentiment analysis methods" are techniques for analyzing user feedback and emotional data and using the results to reduce bias.

[0384] A "dynamic adjustment method" is a means of dynamically adjusting specific content to match user emotions and local culture based on analysis results.

[0385] This invention is a system that takes into account the cultural background of a specific region and the emotions of users to reduce bias in advertising content and improve the user experience. The system mainly consists of data collection means, analysis means, emotion analysis means, dynamic adjustment means, feedback mechanism, and review mechanism.

[0386] The server first uses data collection tools to acquire a dataset containing cultural and social information about a specific region. This dataset is collected from open databases on the internet and information provided by companies. The acquired dataset is then analyzed using analytical tools to identify patterns of bias. This process utilizes data analysis platforms and machine learning algorithms.

[0387] Users view advertising content via their devices, and their emotional data is analyzed using sentiment analysis tools. Sentiment analysis APIs such as IBM Watson and Google Cloud Natural Language API are used, and the results are reflected in the analysis report. Based on these results, the server uses dynamic adjustment tools to adjust the advertising content in real time. This adjustment is based on information stored in the Firebase database.

[0388] Furthermore, the server provides standards for reducing bias through dissemination materials and sends information to companies that have been given certification marks to promote trustworthiness. The feedback mechanism collects user feedback and uses it in conjunction with sentiment analysis to achieve less biased output throughout the system.

[0389] A concrete example is local event advertisements. If a user sees an ad and shows no interest, the system can identify the reason through sentiment analysis and adjust the ad to, for example, family-friendly events or content that might interest them.

[0390] Example of a prompt:

[0391] "How can we emotionally analyze user feedback on ads they've watched and tailor ad content to fit specific regional cultures?"

[0392] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0393] Step 1:

[0394] The server uses data collection methods to acquire datasets containing information specific to a particular region. It investigates information sources such as open databases on the internet and those provided by companies, and aggregates this specific information. The input is regional cultural and social data, and the output is the storage of this data in a database.

[0395] Step 2:

[0396] The server analyzes the acquired dataset using analytical tools and identifies bias patterns. A machine learning algorithm is applied to each dataset to evaluate the presence of bias. The input is the dataset collected in step 1, and the output is data showing bias patterns. This result is used for adjustment in subsequent processing.

[0397] Step 3:

[0398] Users view advertising content via their devices, and the sentiment analysis system collects and analyzes the user's emotional data in response. Specifically, inputs such as the user's reactions and tone of voice during ad display are passed to the sentiment API and output as an emotion score. This analysis result is data that indicates the user's emotional state.

[0399] Step 4:

[0400] The server uses sentiment analysis results to dynamically adjust advertising content in real time. The input is the sentiment score from step 3, and the output is the generation of appropriate content using a generative AI model, which is then sent to the device. For example, it automatically changes the content to be more engaging.

[0401] Step 5:

[0402] Users provide feedback using their devices, which a feedback mechanism collects and analyzes. The output is the analysis of the feedback input, with suggestions for improvement to reduce bias sent to the server. This output is then used to further refine the system's recognition algorithm.

[0403] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0404] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0405] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0406] [Third Embodiment]

[0407] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0408] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0409] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0410] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0411] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0412] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0413] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0414] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0415] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0416] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0417] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0418] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0419] This invention relates to a method and apparatus for constructing and operating an artificial intelligence system that takes into account the specific information of a particular region. This system consists of a server, a terminal, and a user, and each component works in cooperation with the others.

[0420] First, the server collects data from various companies and organizations. This data includes social and cultural information specific to a particular region and is collected using data collection methods. Next, the server uses analytical tools to identify biases within the data. During the analysis process, statistical and machine learning algorithms are used to find specific patterns and correlations, with a particular focus on bias analysis.

[0421] Based on the analysis results, a feedback mechanism is designed. This mechanism collects feedback from users via a terminal and works in conjunction with the analysis means to evaluate bias reduction. The terminal provides an interface that allows users to easily provide feedback. The collected feedback is then used for further analysis on the server.

[0422] Furthermore, the server uses readily available resources to provide companies with best practices and guidelines to mitigate bias. Based on this information, terminals can improve their AI models and take measures to enhance fairness.

[0423] Furthermore, companies that have successfully implemented appropriate bias reduction measures will be granted a certification mark by the certification body. This occurs after users evaluate their own AI systems, and the server assesses whether the system meets the certification standards based on that evaluation. This certification mark serves as a symbol for users to demonstrate their trustworthiness to consumers.

[0424] To give a concrete example, suppose a user provides past misrecognition data regarding specific attributes to mitigate bias in image recognition AI within a particular cultural sphere. Through this feedback, the device updates the model, the server verifies that the improvements are appropriate, and then issues a certification mark. This process is expected to improve the fairness and reliability of the AI.

[0425] The following describes the processing flow.

[0426] Step 1:

[0427] The server collects region-specific information from various companies and organizations. The server uses data collection methods to obtain cultural and social datasets for the specific region.

[0428] Step 2:

[0429] The server processes the collected dataset using analytical tools to identify biases and patterns. The server applies machine learning algorithms for bias analysis to recognize specific biases and imbalances.

[0430] Step 3:

[0431] The server designs a feedback mechanism based on the analysis results. The terminal builds an interface that allows users to easily input feedback.

[0432] Step 4:

[0433] Users provide feedback on the AI ​​system's output via their devices. Users comment on and evaluate specific patterns or questionable results.

[0434] Step 5:

[0435] The server re-evaluates the collected feedback using analytical tools. The server recalculates the degree of bias using the feedback data and identifies areas for improvement in the AI ​​model.

[0436] Step 6:

[0437] The device provides best practices and guidelines for reducing bias using dissemination materials. The device distributes instructional information to users for improvement.

[0438] Step 7:

[0439] Users update or improve their AI models according to the provided guidelines. Users implement specific changes to reduce bias.

[0440] Step 8:

[0441] The server evaluates the improved model with a review body and determines whether it is suitable for certification. The server performs the evaluation based on the certification standards and records the results.

[0442] Step 9:

[0443] The server assigns a certification mark to companies that it deems to have adequately reduced bias. The server then notifies the user of the result and provides it as reliability information.

[0444] (Example 1)

[0445] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0446] Artificial intelligence systems that take into account the cultural diversity of specific domains are highly susceptible to bias. Therefore, there is a need to build fair and reliable systems for specific regions and situations, but current technology makes this difficult to achieve. Furthermore, there is a lack of evaluation criteria to mitigate bias and methods to promote reliability, necessitating a comprehensive solution.

[0447] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0448] In this invention, the server includes means for acquiring an information set containing information specific to a particular domain using an information gathering device, means for identifying biases included in the information set using an analysis device, and means for collecting opinions from users via voice and text input through a dialogue device and performing evaluations for bias reduction. This makes it possible to efficiently collect information specific to a particular domain, identify biases, and appropriately evaluate and reduce them.

[0449] An "information gathering device" is a mechanism for acquiring data, including cultural and social information, related to a specific domain.

[0450] An "information set" is a collection of cultural and social data related to a specific domain, acquired by an information gathering device.

[0451] An "analytical device" is a technical device used to analyze data contained in an information set and identify bias.

[0452] "Bias" refers to an imbalance or bias present in data, indicating a lack of fairness regarding specific attributes or patterns.

[0453] A "dialogue device" is a mechanism that collects opinions and feedback from users through voice or text input and provides data for bias evaluation.

[0454] "Users" refer to those who provide opinions and feedback through dialogue devices and are actively involved in the biased evaluation process.

[0455] "Dissemination materials" are informational materials that present best practices and guidelines for reducing bias and are intended to be disseminated to users.

[0456] "Guidelines" refer to standards and procedures designed to mitigate bias and maintain and improve the fairness of the system.

[0457] A "review body" refers to an organization or process that evaluates the appropriate use of a service in relation to the goal of reducing bias and issues approval marks.

[0458] An "approval mark" indicates that a product has been certified by an accreditation body as being used appropriately, and serves as proof of its reliability and fairness.

[0459] One embodiment of the present invention is to consider cultural and social information related to a specific domain, mitigate bias through an artificial intelligence system, and achieve fair data processing. This system consists of a server, terminals, and users working together.

[0460] The server first uses an information gathering device to obtain specific information about a particular domain from various sources. This device plays a role in continuously collecting data through APIs and external database connections. The server can efficiently acquire data using the Python "requests" library.

[0461] The acquired information is stored as an information set and analyzed via an analysis device on the server. The analysis includes data preprocessing and bias identification, utilizing statistical analysis tools such as "pandas" and "scikit-learn," as well as machine learning libraries. This makes it possible to detect imbalanced patterns within the data and identify bias.

[0462] Users provide opinions based on their everyday experiences and observations using a dialogue device via their terminal. This feedback is submitted via voice or text input and sent to the server through review forms or mobile apps. The terminal then passes the collected opinions to an analysis device to help evaluate and mitigate bias.

[0463] Based on the collected and analyzed data, the server generates guidelines and best practices for bias reduction as dissemination materials. These materials are provided through a portal site accessible to companies and organizations.

[0464] An example of a feedback prompt is, "Please provide your feedback on how we can improve the misrecognition of image recognition AI in a specific social group." This prompt elicits useful opinions based on users' real-world experiences and is used for advanced data analysis.

[0465] Furthermore, if proper use is confirmed, the server will be granted an approval mark through a review mechanism and will transmit information to promote its reliability. This will enable the system to be positioned as fair and reliable for consumers and related organizations.

[0466] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0467] Step 1: Data Collection

[0468] The server utilizes information gathering devices to collect data related to a specific domain. Inputs include information from APIs and external databases, and the server retrieves this data using libraries such as "requests". The output is a dataset encompassing cultural and social information within the specific domain. In this step, a script is executed periodically to retrieve the latest data.

[0469] Step 2: Data Analysis

[0470] The server analyzes the collected dataset using an analytical instrument. The input for this step is the dataset collected in step 1. The server uses the "pandas" and "scikit-learn" libraries to identify patterns and biases in the data. By applying statistical methods, it is possible to verify the integrity of the data and identify biases. The output is an analysis result that identifies biases.

[0471] Step 3: Gathering Feedback

[0472] The terminal uses a dialogue device to collect opinions from the user. The input consists of user opinions in both voice and text format. The terminal collects this feedback through forms and applications and sends it to the server. The output is the collected user feedback data.

[0473] Step 4: Feedback analysis and model update

[0474] The server updates the AI ​​model based on collected feedback to mitigate bias. The input consists of feedback data and previously obtained data analysis results. The generated AI model is retrained using frameworks such as "TensorFlow" or "PyTorch" to update it to a model that takes bias into account. The output is the improved AI model.

[0475] Step 5: Generating and distributing guidelines

[0476] The server generates guidelines for bias reduction based on the analysis results and updated models. The input consists of the analysis results and improvements to the model. The generated guidelines are output as Markdown or PDF files, and this information is distributed to companies and organizations through a dedicated portal.

[0477] Step 6: Evaluation and Authentication

[0478] The server uses an evaluation system to verify the appropriateness of bias mitigation and issues approval marks as needed. Inputs include evaluation criteria, feedback from companies, and analysis reports. After the review process, eligible users can obtain approval marks. Outputs include approval marks and information to promote reliability.

[0479] (Application Example 1)

[0480] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0481] There is a need for methods to mitigate bias in artificial intelligence systems that take into account regionally differing social and cultural backgrounds, thereby improving fairness and reliability. In particular, in technologies closely related to daily life, such as household robots, there is a need to improve the situation as neglecting regional characteristics can lead to model malfunctions and a decline in the quality of the user experience.

[0482] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0483] In this invention, the server includes means for collecting a data set containing attribute information of a specific region using a data acquisition device, means for identifying biases included in the data set using an analysis device, and means for collecting feedback through a user interface and evaluating bias correction that takes into account the diversity of the specific region. This makes it possible to provide an impartial and reliable artificial intelligence system adapted to a specific region.

[0484] A "data acquisition device" is a device for efficiently collecting data sets that include attribute information of a specific region.

[0485] An "analysis device" is a device used to identify biases in a collected data set and to analyze patterns within the data using statistical and machine learning algorithms.

[0486] A "user interface" is an interactive means of collecting feedback from users and evaluating bias corrections that take into account the diversity of a specific region.

[0487] A "learning algorithm" is a set of computational methods and programs used to continuously modify and adjust a machine learning model based on collected feedback data.

[0488] A "distribution device" is a device that provides users and related organizations with guidelines and standard information for proper bias correction.

[0489] An "evaluation device" is a device that evaluates proper operation based on bias correction guidelines and assigns a confidence mark if the standards are met.

[0490] This invention constitutes a region-adaptive artificial intelligence system specifically for household robots. The server efficiently collects region-specific attribute information using a data acquisition device. This data reflects the cultural and social background of a particular region, and bias is identified by an analysis device based on the collected data set. Machine learning libraries such as TensorFlow are used to identify bias, and a model is created to mitigate the bias based on the obtained data.

[0491] The user's device collects feedback through the user interface, and this feedback is sent to the server. The server uses this data to modify the model using a learning algorithm. In particular, corrections are made that take into account individual cultural elements and regional characteristics based on the feedback. Analysis libraries such as scikit-learn are used in this process. The corrected model is then applied by the distribution device to provide highly accurate, region-specific services.

[0492] As a concrete example, a robot designed for households in Tokyo could gather feedback from users about their cleaning methods and supplement its knowledge with specialized information on how to care for tatami mats. This would allow users to receive high-quality, locally tailored services.

[0493] As an example of a specific prompt for the generating AI model, input would be: "Generate cleaning guidelines that a household robot in Tokyo should follow. Pay attention to how to care for tatami mats." This sentence will serve as the basis for providing region-specific instructions to the AI ​​model.

[0494] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0495] Step 1:

[0496] The server collects regional attribute information using data acquisition devices. It receives social and cultural data related to a specific region as input and generates a data set containing regional characteristics as output. Specifically, it aggregates data provided by local businesses and publicly available statistical data.

[0497] Step 2:

[0498] The server identifies bias within a data set collected using an analysis device. It receives the data set as input and performs data analysis using TensorFlow to identify bias. The output generates a report detailing the degree of bias and the types of bias identified. Specific operations include feature extraction and correlation analysis.

[0499] Step 3:

[0500] The user's device collects feedback using a user interface. It acquires feedback data based on user opinions and improvement requests as input. It creates a feedback dataset to be sent to the server as output. Specific operations include data input in the form of questionnaires and opinion collection using speech recognition.

[0501] Step 4:

[0502] The server modifies the machine learning model using a learning algorithm based on feedback data. It receives bias analysis results and user feedback data as input and generates a modified AI model as output. Specifically, it performs data retraining and model parameter adjustment using scikit-learn.

[0503] Step 5:

[0504] The server distributes the corrected model to the terminal via a distribution device and applies it. It receives the corrected AI model as input and reflects the optimized model in the terminal as output. Specifically, it provides operational guidelines based on newly learned processes and updates the model within the terminal.

[0505] Step 6:

[0506] Users will experience region-specific services suggested by robots based on improved AI models and provide feedback on their evaluation. The expected output is improved efficiency of region-specific robot services and increased user satisfaction. Specific examples of operation include the robot suggesting methods for maintaining tatami mats and the user verifying the results after application.

[0507] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0508] This invention relates to a method and apparatus for enhancing the bias reduction process in an artificial intelligence system that takes into account specific regional information, by also analyzing the user's emotions. The system consists of a server, a terminal, a user, and an emotion engine.

[0509] First, the server collects data provided by various companies and organizations. This includes cultural and social information specific to a particular region, and is efficiently acquired through data collection methods. Next, the server uses analysis tools to identify biases within the data and detect specific patterns. In this process, in addition to traditional machine learning algorithms, an emotion engine is used to analyze the sentiment of feedback collected from users.

[0510] The emotion engine operates through a process of processing user feedback and analyzing the emotions it captures. For example, if a user expresses dissatisfaction with an image recognition result, the emotion engine examines the cause of that emotion, and this is taken into consideration in the feedback mechanism. This leads to more precise recognition of biases, and the feedback mechanism is used to implement improvements.

[0511] The terminal provides an interface for users to input feedback. The feedback entered through the user interface is analyzed by an emotion engine, sent to a server, and used for bias assessment and improvement.

[0512] Furthermore, the server provides guidelines and tools to mitigate bias using readily available materials. This information is distributed to companies via terminals, promoting the proper improvement of AI models. In addition, a review body evaluates the improved AI models and issues a certification mark if they meet the standards.

[0513] A concrete example is an AI that generates marketing images for consumers in a specific region. This system evaluates both the user's feedback regarding bias in the image content and the user's emotions simultaneously. The server uses this information to adjust the image generation algorithm, producing more culturally appropriate output. As a result, user satisfaction is expected to improve, and trust will be strengthened.

[0514] The following describes the processing flow.

[0515] Step 1:

[0516] The server collects data provided by companies and organizations through data collection methods. This includes datasets that contain cultural background and social information specific to a particular region.

[0517] Step 2:

[0518] The server processes the collected data using analytical tools to identify any biases or patterns present in the data. The analysis results are recorded in a database on the server.

[0519] Step 3:

[0520] Users input feedback on the AI ​​system's output using the terminal's user interface. This feedback includes specific comments and suggestions for improvement.

[0521] Step 4:

[0522] The device sends user feedback to the emotion engine. The emotion engine analyzes the user's emotions contained in the feedback and determines the type and intensity of those emotions.

[0523] Step 5:

[0524] The server adjusts the feedback mechanism using the analysis results from the emotion engine. In particular, it evaluates areas where strong biases were identified and identifies necessary improvements.

[0525] Step 6:

[0526] The device receives sentiment analysis results from the server and presents the user with improvement measures and suggestions based on the feedback. It also provides guidelines for reducing bias through promotional materials.

[0527] Step 7:

[0528] Users review the suggested improvements via their devices and implement them into their own AI model. They adjust the model parameters as needed.

[0529] Step 8:

[0530] The server activates a review mechanism to evaluate the AI ​​model improved by the user. The evaluation is conducted based on bias mitigation criteria, and if it passes, a certification mark is issued.

[0531] Step 9:

[0532] The device notifies the user that it has been assigned an authentication mark and provides information informing them of the improved reliability of the authenticated AI system.

[0533] (Example 2)

[0534] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0535] There is a need to effectively identify and mitigate biases in datasets within specific regions. However, conventional technologies struggle to adequately consider regional cultural and emotional factors in their bias mitigation. Furthermore, the lack of mechanisms for effectively incorporating user feedback limits model improvement.

[0536] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0537] In this invention, the server includes means for acquiring an information set containing information about local culture and society using a data processing device, means for detecting biases contained in the information set using an analysis device, and means for identifying the emotions of feedback provided by users using an emotion analysis device and reflecting this in the bias detection results. This makes it possible to make appropriate improvements that take into account region-specific biases.

[0538] A "data processing device" is a device for efficiently collecting information about local culture and society and obtaining it as an information set.

[0539] An "analytical device" is a device that detects data contained in an information set and performs analysis to identify biases.

[0540] An "emotion analysis device" is a device that examines the emotions in feedback provided by users and incorporates them into the bias detection process.

[0541] A "user interface" is a system component that provides an interactive means for users to provide feedback.

[0542] An "information distribution device" is a device that generates standards and tools for bias correction and distributes information in order to disseminate them.

[0543] An "evaluation device" is a device that examines whether an improved information processing method conforms to the standards and, if it does, issues a certification mark.

[0544] A "certification mark" is a mark that indicates reliability, given to an information processing method or user that conforms to standards.

[0545] In this invention, a data processing system is constructed through the cooperation of a server, a terminal, and an emotion analysis engine. The main components include a data processing device, an analysis device, an emotion analysis device, a user interface, an information distribution device, and an evaluation device.

[0546] The server uses data processing equipment to collect local cultural and social information and obtain a collection of that information. This allows the server to accumulate diverse information and form a foundation for mitigating bias. Next, the server uses analytical equipment to detect biases in the collected information collection and extract specific patterns and trends. This analysis takes into account the cultural and social context of the specific region to minimize errors.

[0547] The sentiment analyzer is responsible for processing user feedback and analyzing the emotions associated with it. When a user inputs feedback using a terminal, the sentiment analyzer emotionally evaluates the content and incorporates this into bias detection. This analysis is performed to comprehensively understand user satisfaction and dissatisfaction. The user interface provides users with intuitive and easy-to-use means for collecting and inputting feedback.

[0548] Furthermore, the server uses an information distribution device to generate guidelines and tools for reducing bias and distribute them to companies and related organizations. This distribution supports the dissemination of improved information processing methods. The evaluation device checks whether the improved information processing methods conform to predetermined standards. If they conform to the standards, the server assigns a certification mark, indicating improved reliability.

[0549] A concrete example is an AI system that generates marketing images targeted at consumers in a specific region. If a user provides feedback on biases related to an image, the prompt might be something like, "Please enter information to identify and improve biases in marketing images for a specific region." The server can then use this information to adjust its algorithm and obtain a more culturally appropriate generating AI model. As a result, user satisfaction and reliability are expected to improve.

[0550] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0551] Step 1:

[0552] The server uses data processing equipment to collect local cultural and social information. The input is diverse information provided by companies and organizations, and the output is a collection of that information. In this step, the server retrieves data directly from the information sources and stores it in a database, preparing the foundational data necessary for the next analysis.

[0553] Step 2:

[0554] The server analyzes the data set collected using the analysis device. The input is the data set obtained in step 1, and the output is the identification result that detects and identifies bias. Here, the server applies a machine learning algorithm to extract anti-biased patterns in the data. This analysis is performed using statistical methods.

[0555] Step 3:

[0556] The user provides feedback using the device's user interface. The input is the user's feedback, and the output is the feedback data. In this step, the device's interface provides a user-friendly feedback form, allowing users to easily input their opinions and feelings.

[0557] Step 4:

[0558] The emotion analysis device analyzes the feedback received from the user and performs an emotional evaluation. The input is the feedback data obtained in step 3, and the output is the emotional evaluation result. The emotion analysis device uses natural language processing technology to emotionally interpret the feedback and sends the evaluation result as data to the server.

[0559] Step 5:

[0560] The server re-evaluates the bias in the information set based on the data obtained from sentiment analysis. The input is the bias identification result from step 2 and the sentiment evaluation result from step 4, and the output is the re-evaluated bias data. Here, the server combines the sentiment data with the existing bias data and performs bias adjustment. The sentiment element of feedback is incorporated into the re-evaluation process.

[0561] Step 6:

[0562] The server uses an information distribution device to create guidelines for bias reduction and distributes them to companies and related organizations. The input is re-evaluated bias data, and the output is specific improvement guidelines and tools. The server distributes this through digital media, providing the necessary guidance to widely promote the improvement of AI models.

[0563] Step 7:

[0564] The evaluation device checks whether the improved AI model meets the standards. The input is the improved model and its usage results, and the output is the certification result of compliance with the standards. Here, it is confirmed whether the model's accuracy and reliability meet the standards, and if it does, a certification mark is issued. This evaluation ensures that the model can be used with confidence.

[0565] (Application Example 2)

[0566] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the headset-type terminal 314 will be referred to as the "terminal."

[0567] There is a need to consider the diverse cultural backgrounds of specific regions and mitigate bias in content such as advertising. However, conventional systems have struggled to adequately consider regional emotional factors, making it difficult to improve user satisfaction. Furthermore, they lacked the ability to dynamically adjust content, making immediate responses difficult.

[0568] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0569] In this invention, the server includes means for acquiring a dataset containing specific regional information using data collection means, means for identifying biases contained in the dataset using analysis means, means for analyzing user sentiment data using sentiment analysis means and using the results to mitigate bias, and means for dynamically changing content based on the analysis results using dynamic adjustment means. This makes it possible to mitigate bias in a way that is appropriate for the local culture and to adjust content in real time according to the user's emotions.

[0570] "Data collection means" refers to devices and methods for efficiently acquiring datasets containing cultural and social information of a specific region.

[0571] "Analysis means" refers to processes or devices for identifying biases within an acquired dataset and detecting those patterns.

[0572] A "feedback mechanism" is a device or method designed to collect user feedback and use that feedback to reduce bias.

[0573] "Dissemination materials" are means of providing and spreading standards and tools to reduce bias.

[0574] A "certification body" refers to a device or method for evaluating and certifying appropriate use based on bias mitigation criteria.

[0575] "Sentiment analysis methods" are techniques for analyzing user feedback and emotional data and using the results to reduce bias.

[0576] A "dynamic adjustment method" is a means of dynamically adjusting specific content to match user emotions and local culture based on analysis results.

[0577] This invention is a system that takes into account the cultural background of a specific region and the emotions of users to reduce bias in advertising content and improve the user experience. The system mainly consists of data collection means, analysis means, emotion analysis means, dynamic adjustment means, feedback mechanism, and review mechanism.

[0578] The server first uses data collection tools to acquire a dataset containing cultural and social information about a specific region. This dataset is collected from open databases on the internet and information provided by companies. The acquired dataset is then analyzed using analytical tools to identify patterns of bias. This process utilizes data analysis platforms and machine learning algorithms.

[0579] Users view advertising content via their devices, and their emotional data is analyzed using sentiment analysis tools. Sentiment analysis APIs such as IBM Watson and Google Cloud Natural Language API are used, and the results are reflected in the analysis report. Based on these results, the server uses dynamic adjustment tools to adjust the advertising content in real time. This adjustment is based on information stored in the Firebase database.

[0580] Furthermore, the server provides standards for reducing bias through dissemination materials and sends information to companies that have been given certification marks to promote trustworthiness. The feedback mechanism collects user feedback and uses it in conjunction with sentiment analysis to achieve less biased output throughout the system.

[0581] A concrete example is local event advertisements. If a user sees an ad and shows no interest, the system can identify the reason through sentiment analysis and adjust the ad to, for example, family-friendly events or content that might interest them.

[0582] Example of a prompt:

[0583] "How can we emotionally analyze user feedback on ads they've watched and tailor ad content to fit specific regional cultures?"

[0584] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0585] Step 1:

[0586] The server uses data collection methods to acquire datasets containing information specific to a particular region. It investigates information sources such as open databases on the internet and those provided by companies, and aggregates this specific information. The input is regional cultural and social data, and the output is the storage of this data in a database.

[0587] Step 2:

[0588] The server analyzes the acquired dataset using analytical tools and identifies bias patterns. A machine learning algorithm is applied to each dataset to evaluate the presence of bias. The input is the dataset collected in step 1, and the output is data showing bias patterns. This result is used for adjustment in subsequent processing.

[0589] Step 3:

[0590] Users view advertising content via their devices, and the sentiment analysis system collects and analyzes the user's emotional data in response. Specifically, inputs such as the user's reactions and tone of voice during ad display are passed to the sentiment API and output as an emotion score. This analysis result is data that indicates the user's emotional state.

[0591] Step 4:

[0592] The server uses sentiment analysis results to dynamically adjust advertising content in real time. The input is the sentiment score from step 3, and the output is the generation of appropriate content using a generative AI model, which is then sent to the device. For example, it automatically changes the content to be more engaging.

[0593] Step 5:

[0594] Users provide feedback using their devices, which a feedback mechanism collects and analyzes. The output is the analysis of the feedback input, with suggestions for improvement to reduce bias sent to the server. This output is then used to further refine the system's recognition algorithm.

[0595] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0596] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0597] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0598] [Fourth Embodiment]

[0599] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0600] As shown in Figure 7, the data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[0601] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0602] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0603] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0604] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0605] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0606] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0607] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0608] The specific processing program 56 is an example of a "program" relating to the technology of this disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0609] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0610] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0611] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0612] This invention relates to a method and apparatus for constructing and operating an artificial intelligence system that takes into account the specific information of a particular region. This system consists of a server, a terminal, and a user, and each component works in cooperation with the others.

[0613] First, the server collects data from various companies and organizations. This data includes social and cultural information specific to a particular region and is collected using data collection methods. Next, the server uses analytical tools to identify biases within the data. During the analysis process, statistical and machine learning algorithms are used to find specific patterns and correlations, with a particular focus on bias analysis.

[0614] Based on the analysis results, a feedback mechanism is designed. This mechanism collects feedback from users via a terminal and works in conjunction with the analysis means to evaluate bias reduction. The terminal provides an interface that allows users to easily provide feedback. The collected feedback is then used for further analysis on the server.

[0615] Furthermore, the server uses readily available resources to provide companies with best practices and guidelines to mitigate bias. Based on this information, terminals can improve their AI models and take measures to enhance fairness.

[0616] Furthermore, companies that have successfully implemented appropriate bias reduction measures will be granted a certification mark by the certification body. This occurs after users evaluate their own AI systems, and the server assesses whether the system meets the certification standards based on that evaluation. This certification mark serves as a symbol for users to demonstrate their trustworthiness to consumers.

[0617] To give a concrete example, suppose a user provides past misrecognition data regarding specific attributes to mitigate bias in image recognition AI within a particular cultural sphere. Through this feedback, the device updates the model, the server verifies that the improvements are appropriate, and then issues a certification mark. This process is expected to improve the fairness and reliability of the AI.

[0618] The following describes the processing flow.

[0619] Step 1:

[0620] The server collects region-specific information from various companies and organizations. The server uses data collection methods to obtain cultural and social datasets for the specific region.

[0621] Step 2:

[0622] The server processes the collected dataset using analytical tools to identify biases and patterns. The server applies machine learning algorithms for bias analysis to recognize specific biases and imbalances.

[0623] Step 3:

[0624] The server designs a feedback mechanism based on the analysis results. The terminal builds an interface that allows users to easily input feedback.

[0625] Step 4:

[0626] Users provide feedback on the AI ​​system's output via their devices. Users comment on and evaluate specific patterns or questionable results.

[0627] Step 5:

[0628] The server re-evaluates the collected feedback using analytical tools. The server recalculates the degree of bias using the feedback data and identifies areas for improvement in the AI ​​model.

[0629] Step 6:

[0630] The device provides best practices and guidelines for reducing bias using dissemination materials. The device distributes instructional information to users for improvement.

[0631] Step 7:

[0632] Users update or improve their AI models according to the provided guidelines. Users implement specific changes to reduce bias.

[0633] Step 8:

[0634] The server evaluates the improved model with a review body and determines whether it is suitable for certification. The server performs the evaluation based on the certification standards and records the results.

[0635] Step 9:

[0636] The server assigns a certification mark to companies that it deems to have adequately reduced bias. The server then notifies the user of the result and provides it as reliability information.

[0637] (Example 1)

[0638] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0639] Artificial intelligence systems that take into account the cultural diversity of specific domains are highly susceptible to bias. Therefore, there is a need to build fair and reliable systems for specific regions and situations, but current technology makes this difficult to achieve. Furthermore, there is a lack of evaluation criteria to mitigate bias and methods to promote reliability, necessitating a comprehensive solution.

[0640] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0641] In this invention, the server includes means for acquiring an information set containing information specific to a particular domain using an information gathering device, means for identifying biases included in the information set using an analysis device, and means for collecting opinions from users via voice and text input through a dialogue device and performing evaluations for bias reduction. This makes it possible to efficiently collect information specific to a particular domain, identify biases, and appropriately evaluate and reduce them.

[0642] An "information gathering device" is a mechanism for acquiring data, including cultural and social information, related to a specific domain.

[0643] An "information set" is a collection of cultural and social data related to a specific domain, acquired by an information gathering device.

[0644] An "analytical device" is a technical device used to analyze data contained in an information set and identify bias.

[0645] "Bias" refers to an imbalance or bias present in data, indicating a lack of fairness regarding specific attributes or patterns.

[0646] A "dialogue device" is a mechanism that collects opinions and feedback from users through voice or text input and provides data for bias evaluation.

[0647] "Users" refer to those who provide opinions and feedback through dialogue devices and are actively involved in the biased evaluation process.

[0648] "Dissemination materials" are informational materials that present best practices and guidelines for reducing bias and are intended to be disseminated to users.

[0649] "Guidelines" refer to standards and procedures designed to mitigate bias and maintain and improve the fairness of the system.

[0650] A "review body" refers to an organization or process that evaluates the appropriate use of a service in relation to the goal of reducing bias and issues approval marks.

[0651] An "approval mark" indicates that a product has been certified by an accreditation body as being used appropriately, and serves as proof of its reliability and fairness.

[0652] One embodiment of the present invention is to consider cultural and social information related to a specific domain, mitigate bias through an artificial intelligence system, and achieve fair data processing. This system consists of a server, terminals, and users working together.

[0653] The server first uses an information gathering device to obtain specific information about a particular domain from various sources. This device plays a role in continuously collecting data through APIs and external database connections. The server can efficiently acquire data using the Python "requests" library.

[0654] The acquired information is stored as an information set and analyzed via an analysis device on the server. The analysis includes data preprocessing and bias identification, utilizing statistical analysis tools such as "pandas" and "scikit-learn," as well as machine learning libraries. This makes it possible to detect imbalanced patterns within the data and identify bias.

[0655] Users provide opinions based on their everyday experiences and observations using a dialogue device via their terminal. This feedback is submitted via voice or text input and sent to the server through review forms or mobile apps. The terminal then passes the collected opinions to an analysis device to help evaluate and mitigate bias.

[0656] Based on the collected and analyzed data, the server generates guidelines and best practices for bias reduction as dissemination materials. These materials are provided through a portal site accessible to companies and organizations.

[0657] An example of a feedback prompt is, "Please provide your feedback on how we can improve the misrecognition of image recognition AI in a specific social group." This prompt elicits useful opinions based on users' real-world experiences and is used for advanced data analysis.

[0658] Furthermore, if proper use is confirmed, the server will be granted an approval mark through a review mechanism and will transmit information to promote its reliability. This will enable the system to be positioned as fair and reliable for consumers and related organizations.

[0659] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0660] Step 1: Data Collection

[0661] The server utilizes information gathering devices to collect data related to a specific domain. Inputs include information from APIs and external databases, and the server retrieves this data using libraries such as "requests". The output is a dataset encompassing cultural and social information within the specific domain. In this step, a script is executed periodically to retrieve the latest data.

[0662] Step 2: Data Analysis

[0663] The server analyzes the collected dataset using an analytical instrument. The input for this step is the dataset collected in step 1. The server uses the "pandas" and "scikit-learn" libraries to identify patterns and biases in the data. By applying statistical methods, it is possible to verify the integrity of the data and identify biases. The output is an analysis result that identifies biases.

[0664] Step 3: Gathering Feedback

[0665] The terminal uses a dialogue device to collect opinions from the user. The input consists of user opinions in both voice and text format. The terminal collects this feedback through forms and applications and sends it to the server. The output is the collected user feedback data.

[0666] Step 4: Feedback analysis and model update

[0667] The server updates the AI ​​model based on collected feedback to mitigate bias. The input consists of feedback data and previously obtained data analysis results. The generated AI model is retrained using frameworks such as "TensorFlow" or "PyTorch" to update it to a model that takes bias into account. The output is the improved AI model.

[0668] Step 5: Generating and distributing guidelines

[0669] The server generates guidelines for bias reduction based on the analysis results and updated models. The input consists of the analysis results and improvements to the model. The generated guidelines are output as Markdown or PDF files, and this information is distributed to companies and organizations through a dedicated portal.

[0670] Step 6: Evaluation and Authentication

[0671] The server uses an evaluation system to verify the appropriateness of bias mitigation and issues approval marks as needed. Inputs include evaluation criteria, feedback from companies, and analysis reports. After the review process, eligible users can obtain approval marks. Outputs include approval marks and information to promote reliability.

[0672] (Application Example 1)

[0673] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0674] There is a need for methods to mitigate bias in artificial intelligence systems that take into account regionally differing social and cultural backgrounds, thereby improving fairness and reliability. In particular, in technologies closely related to daily life, such as household robots, there is a need to improve the situation as neglecting regional characteristics can lead to model malfunctions and a decline in the quality of the user experience.

[0675] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0676] In this invention, the server includes means for collecting a data set containing attribute information of a specific region using a data acquisition device, means for identifying biases included in the data set using an analysis device, and means for collecting feedback through a user interface and evaluating bias correction that takes into account the diversity of the specific region. This makes it possible to provide an impartial and reliable artificial intelligence system adapted to a specific region.

[0677] A "data acquisition device" is a device for efficiently collecting data sets that include attribute information of a specific region.

[0678] An "analysis device" is a device used to identify biases in a collected data set and to analyze patterns within the data using statistical and machine learning algorithms.

[0679] A "user interface" is an interactive means of collecting feedback from users and evaluating bias corrections that take into account the diversity of a specific region.

[0680] A "learning algorithm" is a set of computational methods and programs used to continuously modify and adjust a machine learning model based on collected feedback data.

[0681] A "distribution device" is a device that provides users and related organizations with guidelines and standard information for proper bias correction.

[0682] An "evaluation device" is a device that evaluates proper operation based on bias correction guidelines and assigns a confidence mark if the standards are met.

[0683] This invention constitutes a region-adaptive artificial intelligence system specifically for household robots. The server efficiently collects region-specific attribute information using a data acquisition device. This data reflects the cultural and social background of a particular region, and bias is identified by an analysis device based on the collected data set. Machine learning libraries such as TensorFlow are used to identify bias, and a model is created to mitigate the bias based on the obtained data.

[0684] The user's device collects feedback through the user interface, and this feedback is sent to the server. The server uses this data to modify the model using a learning algorithm. In particular, corrections are made that take into account individual cultural elements and regional characteristics based on the feedback. Analysis libraries such as scikit-learn are used in this process. The corrected model is then applied by the distribution device to provide highly accurate, region-specific services.

[0685] As a concrete example, a robot designed for households in Tokyo could gather feedback from users about their cleaning methods and supplement its knowledge with specialized information on how to care for tatami mats. This would allow users to receive high-quality, locally tailored services.

[0686] As an example of a specific prompt for the generating AI model, input would be: "Generate cleaning guidelines that a household robot in Tokyo should follow. Pay attention to how to care for tatami mats." This sentence will serve as the basis for providing region-specific instructions to the AI ​​model.

[0687] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0688] Step 1:

[0689] The server collects regional attribute information using data acquisition devices. It receives social and cultural data related to a specific region as input and generates a data set containing regional characteristics as output. Specifically, it aggregates data provided by local businesses and publicly available statistical data.

[0690] Step 2:

[0691] The server identifies bias within a data set collected using an analysis device. It receives the data set as input and performs data analysis using TensorFlow to identify bias. The output generates a report detailing the degree of bias and the types of bias identified. Specific operations include feature extraction and correlation analysis.

[0692] Step 3:

[0693] The user's device collects feedback using a user interface. It acquires feedback data based on user opinions and improvement requests as input. It creates a feedback dataset to be sent to the server as output. Specific operations include data input in the form of questionnaires and opinion collection using speech recognition.

[0694] Step 4:

[0695] The server modifies the machine learning model using a learning algorithm based on feedback data. It receives bias analysis results and user feedback data as input and generates a modified AI model as output. Specifically, it performs data retraining and model parameter adjustment using scikit-learn.

[0696] Step 5:

[0697] The server distributes the corrected model to the terminal via a distribution device and applies it. It receives the corrected AI model as input and reflects the optimized model in the terminal as output. Specifically, it provides operational guidelines based on newly learned processes and updates the model within the terminal.

[0698] Step 6:

[0699] Users will experience region-specific services suggested by robots based on improved AI models and provide feedback on their evaluation. The expected output is improved efficiency of region-specific robot services and increased user satisfaction. Specific examples of operation include the robot suggesting methods for maintaining tatami mats and the user verifying the results after application.

[0700] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0701] This invention relates to a method and apparatus for enhancing the bias reduction process in an artificial intelligence system that takes into account specific regional information, by also analyzing the user's emotions. The system consists of a server, a terminal, a user, and an emotion engine.

[0702] First, the server collects data provided by various companies and organizations. This includes cultural and social information specific to a particular region, and is efficiently acquired through data collection methods. Next, the server uses analysis tools to identify biases within the data and detect specific patterns. In this process, in addition to traditional machine learning algorithms, an emotion engine is used to analyze the sentiment of feedback collected from users.

[0703] The emotion engine operates through a process of processing user feedback and analyzing the emotions it captures. For example, if a user expresses dissatisfaction with an image recognition result, the emotion engine examines the cause of that emotion, and this is taken into consideration in the feedback mechanism. This leads to more precise recognition of biases, and the feedback mechanism is used to implement improvements.

[0704] The terminal provides an interface for users to input feedback. The feedback entered through the user interface is analyzed by an emotion engine, sent to a server, and used for bias assessment and improvement.

[0705] Furthermore, the server provides guidelines and tools to mitigate bias using readily available materials. This information is distributed to companies via terminals, promoting the proper improvement of AI models. In addition, a review body evaluates the improved AI models and issues a certification mark if they meet the standards.

[0706] A concrete example is an AI that generates marketing images for consumers in a specific region. This system evaluates both the user's feedback regarding bias in the image content and the user's emotions simultaneously. The server uses this information to adjust the image generation algorithm, producing more culturally appropriate output. As a result, user satisfaction is expected to improve, and trust will be strengthened.

[0707] The following describes the processing flow.

[0708] Step 1:

[0709] The server collects data provided by companies and organizations through data collection methods. This includes datasets that contain cultural background and social information specific to a particular region.

[0710] Step 2:

[0711] The server processes the collected data using analytical tools to identify any biases or patterns present in the data. The analysis results are recorded in a database on the server.

[0712] Step 3:

[0713] Users input feedback on the AI ​​system's output using the terminal's user interface. This feedback includes specific comments and suggestions for improvement.

[0714] Step 4:

[0715] The device sends user feedback to the emotion engine. The emotion engine analyzes the user's emotions contained in the feedback and determines the type and intensity of those emotions.

[0716] Step 5:

[0717] The server adjusts the feedback mechanism using the analysis results from the emotion engine. In particular, it evaluates areas where strong biases were identified and identifies necessary improvements.

[0718] Step 6:

[0719] The device receives sentiment analysis results from the server and presents the user with improvement measures and suggestions based on the feedback. It also provides guidelines for reducing bias through promotional materials.

[0720] Step 7:

[0721] Users review the suggested improvements via their devices and implement them into their own AI model. They adjust the model parameters as needed.

[0722] Step 8:

[0723] The server activates a review mechanism to evaluate the AI ​​model improved by the user. The evaluation is conducted based on bias mitigation criteria, and if it passes, a certification mark is issued.

[0724] Step 9:

[0725] The device notifies the user that it has been assigned an authentication mark and provides information informing them of the improved reliability of the authenticated AI system.

[0726] (Example 2)

[0727] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0728] There is a need to effectively identify and mitigate biases in datasets within specific regions. However, conventional technologies struggle to adequately consider regional cultural and emotional factors in their bias mitigation. Furthermore, the lack of mechanisms for effectively incorporating user feedback limits model improvement.

[0729] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0730] In this invention, the server includes means for acquiring an information set containing information about local culture and society using a data processing device, means for detecting biases contained in the information set using an analysis device, and means for identifying the emotions of feedback provided by users using an emotion analysis device and reflecting this in the bias detection results. This makes it possible to make appropriate improvements that take into account region-specific biases.

[0731] A "data processing device" is a device for efficiently collecting information about local culture and society and obtaining it as an information set.

[0732] An "analytical device" is a device that detects data contained in an information set and performs analysis to identify biases.

[0733] An "emotion analysis device" is a device that examines the emotions in feedback provided by users and incorporates them into the bias detection process.

[0734] A "user interface" is a system component that provides an interactive means for users to provide feedback.

[0735] An "information distribution device" is a device that generates standards and tools for bias correction and distributes information in order to disseminate them.

[0736] An "evaluation device" is a device that examines whether an improved information processing method conforms to the standards and, if it does, issues a certification mark.

[0737] A "certification mark" is a mark that indicates reliability, given to an information processing method or user that conforms to standards.

[0738] In this invention, a data processing system is constructed through the cooperation of a server, a terminal, and an emotion analysis engine. The main components include a data processing device, an analysis device, an emotion analysis device, a user interface, an information distribution device, and an evaluation device.

[0739] The server uses data processing equipment to collect local cultural and social information and obtain a collection of that information. This allows the server to accumulate diverse information and form a foundation for mitigating bias. Next, the server uses analytical equipment to detect biases in the collected information collection and extract specific patterns and trends. This analysis takes into account the cultural and social context of the specific region to minimize errors.

[0740] The sentiment analyzer is responsible for processing user feedback and analyzing the emotions associated with it. When a user inputs feedback using a terminal, the sentiment analyzer emotionally evaluates the content and incorporates this into bias detection. This analysis is performed to comprehensively understand user satisfaction and dissatisfaction. The user interface provides users with intuitive and easy-to-use means for collecting and inputting feedback.

[0741] Furthermore, the server uses an information distribution device to generate guidelines and tools for reducing bias and distribute them to companies and related organizations. This distribution supports the dissemination of improved information processing methods. The evaluation device checks whether the improved information processing methods conform to predetermined standards. If they conform to the standards, the server assigns a certification mark, indicating improved reliability.

[0742] A concrete example is an AI system that generates marketing images targeted at consumers in a specific region. If a user provides feedback on biases related to an image, the prompt might be something like, "Please enter information to identify and improve biases in marketing images for a specific region." The server can then use this information to adjust its algorithm and obtain a more culturally appropriate generating AI model. As a result, user satisfaction and reliability are expected to improve.

[0743] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0744] Step 1:

[0745] The server uses data processing equipment to collect local cultural and social information. The input is diverse information provided by companies and organizations, and the output is a collection of that information. In this step, the server retrieves data directly from the information sources and stores it in a database, preparing the foundational data necessary for the next analysis.

[0746] Step 2:

[0747] The server analyzes the data set collected using the analysis device. The input is the data set obtained in step 1, and the output is the identification result that detects and identifies bias. Here, the server applies a machine learning algorithm to extract anti-biased patterns in the data. This analysis is performed using statistical methods.

[0748] Step 3:

[0749] The user provides feedback using the device's user interface. The input is the user's feedback, and the output is the feedback data. In this step, the device's interface provides a user-friendly feedback form, allowing users to easily input their opinions and feelings.

[0750] Step 4:

[0751] The emotion analysis device analyzes the feedback received from the user and performs an emotional evaluation. The input is the feedback data obtained in step 3, and the output is the emotional evaluation result. The emotion analysis device uses natural language processing technology to emotionally interpret the feedback and sends the evaluation result as data to the server.

[0752] Step 5:

[0753] The server re-evaluates the bias in the information set based on the data obtained from sentiment analysis. The input is the bias identification result from step 2 and the sentiment evaluation result from step 4, and the output is the re-evaluated bias data. Here, the server combines the sentiment data with the existing bias data and performs bias adjustment. The sentiment element of feedback is incorporated into the re-evaluation process.

[0754] Step 6:

[0755] The server uses an information distribution device to create guidelines for bias reduction and distributes them to companies and related organizations. The input is re-evaluated bias data, and the output is specific improvement guidelines and tools. The server distributes this through digital media, providing the necessary guidance to widely promote the improvement of AI models.

[0756] Step 7:

[0757] The evaluation device checks whether the improved AI model meets the standards. The input is the improved model and its usage results, and the output is the certification result of compliance with the standards. Here, it is confirmed whether the model's accuracy and reliability meet the standards, and if it does, a certification mark is issued. This evaluation ensures that the model can be used with confidence.

[0758] (Application Example 2)

[0759] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0760] There is a need to consider the diverse cultural backgrounds of specific regions and mitigate bias in content such as advertising. However, conventional systems have struggled to adequately consider regional emotional factors, making it difficult to improve user satisfaction. Furthermore, they lacked the ability to dynamically adjust content, making immediate responses difficult.

[0761] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means.

[0762] In this invention, the server includes means for acquiring a dataset containing specific regional information using data collection means, means for identifying biases contained in the dataset using analysis means, means for analyzing user sentiment data using sentiment analysis means and using the results to mitigate bias, and means for dynamically changing content based on the analysis results using dynamic adjustment means. This makes it possible to mitigate bias in a way that is appropriate for the local culture and to adjust content in real time according to the user's emotions.

[0763] "Data collection means" refers to devices and methods for efficiently acquiring datasets containing cultural and social information of a specific region.

[0764] "Analysis means" refers to processes or devices for identifying biases within an acquired dataset and detecting those patterns.

[0765] A "feedback mechanism" is a device or method designed to collect user feedback and use that feedback to reduce bias.

[0766] "Dissemination materials" are means of providing and spreading standards and tools to reduce bias.

[0767] A "certification body" refers to a device or method for evaluating and certifying appropriate use based on bias mitigation criteria.

[0768] "Sentiment analysis methods" are techniques for analyzing user feedback and emotional data and using the results to reduce bias.

[0769] A "dynamic adjustment method" is a means of dynamically adjusting specific content to match user emotions and local culture based on analysis results.

[0770] This invention is a system that takes into account the cultural background of a specific region and the emotions of users to reduce bias in advertising content and improve the user experience. The system mainly consists of data collection means, analysis means, emotion analysis means, dynamic adjustment means, feedback mechanism, and review mechanism.

[0771] The server first uses data collection tools to acquire a dataset containing cultural and social information about a specific region. This dataset is collected from open databases on the internet and information provided by companies. The acquired dataset is then analyzed using analytical tools to identify patterns of bias. This process utilizes data analysis platforms and machine learning algorithms.

[0772] Users view advertising content via their devices, and their emotional data is analyzed using sentiment analysis tools. Sentiment analysis APIs such as IBM Watson and Google Cloud Natural Language API are used, and the results are reflected in the analysis report. Based on these results, the server uses dynamic adjustment tools to adjust the advertising content in real time. This adjustment is based on information stored in the Firebase database.

[0773] Furthermore, the server provides standards for reducing bias through dissemination materials and sends information to companies that have been given certification marks to promote trustworthiness. The feedback mechanism collects user feedback and uses it in conjunction with sentiment analysis to achieve less biased output throughout the system.

[0774] A concrete example is local event advertisements. If a user sees an ad and shows no interest, the system can identify the reason through sentiment analysis and adjust the ad to, for example, family-friendly events or content that might interest them.

[0775] Example of a prompt:

[0776] "How can we emotionally analyze user feedback on ads they've watched and tailor ad content to fit specific regional cultures?"

[0777] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0778] Step 1:

[0779] The server uses data collection methods to acquire datasets containing information specific to a particular region. It investigates information sources such as open databases on the internet and those provided by companies, and aggregates this specific information. The input is regional cultural and social data, and the output is the storage of this data in a database.

[0780] Step 2:

[0781] The server analyzes the acquired dataset using analytical tools and identifies bias patterns. A machine learning algorithm is applied to each dataset to evaluate the presence of bias. The input is the dataset collected in step 1, and the output is data showing bias patterns. This result is used for adjustment in subsequent processing.

[0782] Step 3:

[0783] Users view advertising content via their devices, and the sentiment analysis system collects and analyzes the user's emotional data in response. Specifically, inputs such as the user's reactions and tone of voice during ad display are passed to the sentiment API and output as an emotion score. This analysis result is data that indicates the user's emotional state.

[0784] Step 4:

[0785] The server uses sentiment analysis results to dynamically adjust advertising content in real time. The input is the sentiment score from step 3, and the output is the generation of appropriate content using a generative AI model, which is then sent to the device. For example, it automatically changes the content to be more engaging.

[0786] Step 5:

[0787] Users provide feedback using their devices, which a feedback mechanism collects and analyzes. The output is the analysis of the feedback input, with suggestions for improvement to reduce bias sent to the server. This output is then used to further refine the system's recognition algorithm.

[0788] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0789] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0790] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[0791] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[0792] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[0793] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[0794] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[0795] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[0796] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[0797] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[0798] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[0799] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[0800] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[0801] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[0802] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[0803] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[0804] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[0805] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[0806] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[0807] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[0808] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted to be incorporated by reference.

[0809] The following is further disclosed regarding the embodiments described above.

[0810] (Claim 1)

[0811] A means for acquiring a dataset containing information specific to a particular region through data collection methods,

[0812] The analysis means includes means for identifying biases included in the dataset,

[0813] A feedback mechanism is designed, and means are used to perform evaluations that take into account the diversity of specific regions in order to mitigate the bias.

[0814] The dissemination of materials provides a means to provide standards for reducing bias,

[0815] A means by which a review body evaluates appropriate use based on bias reduction criteria and issues a certification mark,

[0816] A system that includes this.

[0817] (Claim 2)

[0818] The system according to claim 1, wherein the analysis means evaluates the strengthening or reduction of the bias by collecting feedback via a user interface.

[0819] (Claim 3)

[0820] The system according to claim 1, which provides information to promote the reliability of legitimate users who have been given the aforementioned authentication mark.

[0821] "Example 1"

[0822] (Claim 1)

[0823] A means for acquiring an information collection device that includes information specific to a particular domain,

[0824] The analytical device includes means for identifying biases included in the information set,

[0825] A means of collecting opinions from users via voice and text input through a dialogue device and conducting evaluations to reduce bias,

[0826] Dissemination materials provide a means of providing guidelines for reducing bias,

[0827] The review body conducts an evaluation of appropriate use based on the bias reduction guidelines and grants an approval mark,

[0828] A system that includes this.

[0829] (Claim 2)

[0830] The system according to claim 1, wherein the analysis device evaluates the enhancement or reduction of the bias by collecting opinions via a user terminal.

[0831] (Claim 3)

[0832] The system according to claim 1, which transmits information to promote the reliability of a qualified user who has been given the aforementioned approval mark.

[0833] "Application Example 1"

[0834] (Claim 1)

[0835] A means for collecting a data set containing attribute information of a specific region using a data acquisition device,

[0836] The analysis device includes means for identifying biases included in the data set,

[0837] A means of collecting feedback through a user interface and evaluating bias correction that takes into account the diversity of a specific region,

[0838] A method for modifying a machine learning model based on collected feedback data using a learning algorithm,

[0839] The distribution device provides means for providing guidelines for appropriate bias correction,

[0840] An evaluation device is used to evaluate proper operation based on bias correction guidelines and to assign a reliability mark.

[0841] A system that includes this.

[0842] (Claim 2)

[0843] The system according to claim 1, which provides a regionally adaptive artificial intelligence for optimizing data processing based on the cultural background of the region.

[0844] (Claim 3)

[0845] The system according to claim 1, which provides information to improve the reliability of properly operated operators who have been assigned a trust mark.

[0846] "Example 2 of combining an emotion engine"

[0847] (Claim 1)

[0848] A data processing device provides means for acquiring a set of information including information about local culture and society,

[0849] The analytical device includes means for detecting biases contained in the information set,

[0850] A means for using an emotion analysis device to identify the emotions in the feedback provided by the user and reflecting them in the bias detection results,

[0851] A means for collecting user feedback through a user interface and adjusting the bias using the analysis device,

[0852] The information distribution device provides criteria and tools for bias correction and means for disseminating improved information processing methods.

[0853] An evaluation device is used to evaluate whether the improved information processing method conforms to the standard, and if it does, a certification mark is affixed.

[0854] A system that includes this.

[0855] (Claim 2)

[0856] The system according to claim 1, which uses an emotion analysis device to analyze emotions expressed through feedback, and uses the information processing device to optimize the identification and reduction of bias.

[0857] (Claim 3)

[0858] The system according to claim 1, which distributes information to legitimate users who have been assigned the aforementioned authentication mark in order to improve their reliability.

[0859] "Application example 2 when combining with an emotional engine"

[0860] (Claim 1)

[0861] A means for acquiring a dataset containing information specific to a particular region through data collection methods,

[0862] The analysis means includes means for identifying biases included in the dataset,

[0863] A feedback mechanism is designed, and means are used to perform evaluations that take into account the diversity of specific regions in order to mitigate the bias.

[0864] The dissemination of materials provides a means to provide standards for reducing bias,

[0865] A means by which a review body evaluates appropriate use based on bias reduction criteria and issues a certification mark,

[0866] A means of analyzing user emotional data using emotion analysis methods and using the results to reduce bias,

[0867] Dynamic adjustment means include means for dynamically changing content based on analysis results,

[0868] A system that includes this.

[0869] (Claim 2)

[0870] The system according to claim 1, wherein the analysis means evaluates the strengthening or reduction of the bias by collecting feedback via a user interface.

[0871] (Claim 3)

[0872] The system according to claim 1, which provides information to promote the reliability of legitimate users who have been given the aforementioned authentication mark. [Explanation of symbols]

[0873] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means for acquiring a dataset containing information specific to a particular region through data collection methods, The analysis means includes means for identifying biases included in the dataset, A feedback mechanism is designed, and means are used to perform evaluations that take into account the diversity of specific regions in order to mitigate the bias. The dissemination of materials provides a means to provide standards for reducing bias, A means by which a review body evaluates appropriate use based on bias reduction criteria and issues a certification mark, A system that includes this.

2. The system according to claim 1, wherein the analysis means evaluates the strengthening or reduction of the bias by collecting feedback via a user interface.

3. The system according to claim 1, which provides information to promote the reliability of legitimate users who have been given the aforementioned authentication mark.

Images